Triterpene glycosides of Lupinus angustifolius

Article abstract of DOI:10.1016/S0031-9422(02)00112-7

Investigation of whole seeds of Lupinus angustifolius L. (Leguminosae) yielded the two triterpenoid saponins with branched monosaccharide chain 3β,21β,22β,24- tetrahydroxyolean-12-en-3-O-α-L-rhamnopyranosyl-(1→3)- [α-L-rhamnopyranosyl-(1→2)]- β-D-galactop

Full text of DOI:10.1016/S0031-9422(02)00112-7

Phytochemistry 60 (2002) 323–327
www.elsevier.com/locate/phytochem
Triterpene glycosides of Lupinus angustifolius
Girma M. Woldemichaela,b, Michael Winka,*
aInstitut fur Pharmazeutische Biologie, Universitat Heidelberg, Im Neuenheimer Feld 364, 69120 Heidelberg, Germany
¨
¨
bDepartment of Pharmacognosy, School of Pharmacy, Addis Ababa University, P.O.Box 1176, Addis Ababa, Ethiopia
Received 14 January 2002; received in revised form 14 March 2002
Abstract
Investigation of whole seeds of Lupinus angustifolius L. (Leguminosae) yielded the two triterpenoid saponins with branched
monosaccharide chain 3b,21b,22b,24-tetrahydroxyolean-12-en-3-O-a-l-rhamnopyranosyl-(1!3)-[a-l-rhamnopyranosyl-(1!2)]-
b-d-galactopyranosyl-(1!2)-b-d-glucuronopyranoside (3) and 3b,21b,22b,24-tetrahydroxyolean-12-en-3-O-a-l-rhamnopyranosyl-
(1!3)-[a-l-rhamnopyranosyl-(1!2)]-b-d-galactopyranosyl-(1!2)-b-d-glucuronopyranosyl-21-O-a-l-rhamnopyranoside (4) along
with the known compounds soyasaponin I (1) and 3b,21b,22b,24-tetrahydroxyolean-12-en-3-O-a-l-rhamnopyranosyl-(1!2)-b-d-
galactopyranosyl-(1!2)-b-d-glucuronopyranosyl-21-O-a-l-rhamnopyranoside (2). The structures of the compounds were eluci-
dated using hydrolysis, FAB–MS and extensive NMR experiments. Compounds 2–4 showed moderate antifungal activity against
Candida albicans with MIC values of 25, 25 and 30 mg/ml, respectively. Only soyasaponin I was found weakly hemolytic (HC50
>500 mg/ml). # 2002 Elsevier Science Ltd. All rights reserved.
Keywords: Lupinus angustifolius; Leguminosae; Triterpenoid saponin; Antifungal activity; Soyasapogenol glycosides; Candida albicans
1. Introduction
extracts of debittered lupin seeds (with very low alkaloid
levels) appear to potentiate glucose-induced insulin
release. Thus, the antidiabetic activity could be due to
lupin saponins. Attempts at identification of the sapo-
nins present in the seeds of L. angustifolius based on
their chromatographic properties have so far indicated
that soyasaponins I and VI were present (Ruiz et al.,
1995a,b), but there remained a few triterpene glycoside
constituents that awaited further investigation. This
work was undertaken to identify saponins present in
L. angustifolius. Investigation of antidiabetic activity of
the isolated lupin saponins is currently underway.
As part of an investigation into plants traditionally
used in the preparation of food for people with diabetes,
the genus Lupinus was selected for study. This choice
was based on reports that the seed flour of its various
species is incorporated in marmalades as part of the
diabetic diet in several countries (Kusmenoglu et al.,
1998; Villarroel et al., 1996; Mohamed et al., 1993).
Investigation of any possible antidiabetic effect of lupins
in general has until now mainly focused on the quino-
lizidine alkaloids and a series of alkaloids have been
reported as characteristic constituents and are generally
thought to be responsible for the hypoglycemic activity
of the seed powder. Some reports show that lupin alka-
loids such as lupanine, sparteine, multiflorine and
N-methylcytisine in fact do possess hypoglycemic activ-
ity (Kubo et al., 2000; Mohamed et al., 1993). In addi-
tion to up to 2% alkaloids, however, many lupin seeds
(including those of Lupinus angustifolius) are also
known to contain up to 5% saponins (Wink, 1993,
1998). Pereire et al. (2001) reported that aqueous
2. Results and discussion
Methanolic extract of the seed powder of L. angusti-
folius gave saponins 1, 2, 3 and 4. Saponins 1 and 2 were
identified as soyasaponin I and 3b,21b,22b,24-tetra-
hydroxyolean-12-en 3-O-[a-l-rhamnopyranosyl-(1!2)-
b-d-galactopyranosyl-(1!2)-b-d-glucuronopyranosyl]-
21-O-a-l-rhamnopyranoside based on their FAB–MS
and NMR spectra. These two compounds are already
known to be present in Glycine max (Hosny and
Rosazza, 1999) and Medicago sativa (Bialy et al., 1999),
respectively.
* Corresponding author. Tel.: +49-6221-544880/81; fax: +49-
6221-544884.
E-mail address: wink@uni-hd.de (M. Wink).
0031-9422/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0031-9422(02)00112-7

324
G.M. Woldemichael, M. Wink / Phytochemistry 60 (2002) 323–327
1
respectively. The aglycon moiety region in the H and
13C NMR spectra also showed great similarity to that in
2. Signals for seven tertiary methyl groups (ꢀ 0.65, 0.91,
1.01, 1.28, 1.29, 1.36 and 1.47), one trisubstituted olefin
at ꢀ 5.26, (t-like, H-12), an oxymethine proton at ꢀ 3.37
(dd, J=11.8, 3.7 Hz, H-3), two additional oxymethine
protons which with selective TOCSY excitation showed
coupling only with each other at ꢀ 3.62 (d, J=3.1, H-22)
and 4.07 (d, J=3.1, H-21), and two geminal oxymethy-
lene protons at ꢀ 3.35 (d, J=11.5, H-24a) and 4.32
1
(d, J=11.5, H-24b), could be discerned in the H spec-
trum. Chemical shifts for two olefinic carbons at ꢀ 122.5
(C-12) and 144.3 (C-13), characteristic of a Á12-b-
amyrin skeleton, were also observed in the 13C spec-
trum (Connolly and Hill, 1991). The foregoing data
thus enabled the identification of the aglycon of 3 as
3b,21b,22b,24-tetrahydroxyolean-12-en or soyapogenol
A, a compound already isolated from Pueraria radix
(Arao et al., 1997).
Compound 3 gave a positive Liebermann–Burchard
test. As in 2, C54H88O23 was determined as its molecular
formula based on negative and positive ion HR–FABMS,
which gave [MꢀH]ꢀ at m/z 1103.5560 and [M+Na]+
and [M+2NaꢀH]+ at m/z 1127.5704 and 1149.5493,
On acid hydrolysis, 3 afforded d-glucuronic acid,
d-galactose and l-rhamnose, which were identified by
co-TLC and co-injection as their trimethylsilyl deriva-
tives on GC–MS with authentic samples. However, the
Table 1
13C chemical shift assignments and corresponding 1H NMR correlation data of the aglycon region of compounds 3 and 4 in pyridine-d5
a
Positionb
3
4
ꢀC
ꢀH (J in Hz)
ꢀC
ꢀH (J in Hz)
1 (CH2)
2 (CH2)
3 (CH)
38.5
0.84 m, 1.32 m
1.89 overlapped, 2.32 m
3.37 dd (11.8, 3.7)
38.5
0.81 m, 1.33 m
1.93 m, 2.19 m
3.38 dd (11.9, 3.7)
26.5
91.5
43.6
56.0
18.4
32.8
40.1
47.6
36.6
24.0
122.5
144.3
42.0
26.5
27.4
39.3
43.4
47.0
36.3
74.6
79.3
22.7
63.3
15.5
16.7
26.6
22.0
31.0
21.7
26.4
91.1
43.8
55.9
18.4
32.8
40.1
47.7
36.6
24.0
122.5
144.3
42.0
26.4
27.4
39.3
44.0
47.4
36.8
84.2
78.8
22.9
63.5
15.7
16.7
26.8
22.0
31.2
21.8
4 (C)
5 (CH)
0.88 t (12.8)
1.25 overlapped, 1.57 m
1.25 overlapped, 1.49 m
0.87 t (13.0)
1.25 overlapped, 1.54 m
1.26 overlapped, 1.48 m
6 (CH2)
7 (CH2)
8 (C)
9 (CH)
10 (C)
1.62 d (6.0)
1.60 d (6.1)
11 (CH2)
12 (CH)
13 (C)
1.81 d (6.1)
5.26 t-like
1.80 dd (6.1, 0.7)
5.26 t (0.7)
14 (C)
15 (CH2)
16 (CH2)
17 (C)
0.93 m, 1.86 overlapped
1.10 m, 1.95 m
0.99 overlapped, 1.83 m
1.09 m, 1.95 m
18 (CH)
19 (CH2)
20 (C)
2.37 dd (13.1, 2.0)
1.44 t (13.1), 1.52 dd (13.1, 2.0)
2.53 dd (13.5, 1.9)
1.26 dd (13.5, 1.9), 2.01 t (13.5)
21 (CH)
22 (CH)
23 (CH3)
24 (CH2)
25 (CH3)
26 (CH3)
27 (CH3)
28 (CH3)
29 (CH3)
30 (CH3)
4.07 d (3.1)
3.62 d (3.1)
3.82 d (2.8)
3.91 d (2.8)
1.47 s
3.35 d (11.5), 4.32 d (11.5)
1.44 s
3.26 d (11.5), 4.26 d (11.5)
0.65 s
0.91 s
1.29 s
1.28 s
1.01 s
1.36 s
0.67 s
0.92 s
1.29 s
1.27 s
1.01 s
1.36 s
a
Assignments based on 1H, 13C, DEPT, HSQC, HMBC and selective 1D-TOCSY experiments.
Multiplicities as determined by DEPT.
b

G.M. Woldemichael, M. Wink / Phytochemistry 60 (2002) 323–327
325
1H NMR spectrum showed four anomeric protons at
ꢀ 5.02, 5.86, 6.26 and 6.38. The corresponding anomeric
carbons were located at ꢀ 105.5, 102.1, 102.7 and 102.3
using their HMQC correlations. This suggested, as in 2,
the presence of two l-rhamnose residues. Confirmation
was obtained by the appearance of a signal for a methyl
doublet at ꢀ 1.80 (6H, d, J=6.8 Hz). The site of attach-
ment and the sequence in the monosaccharide chain was
determined through the analysis of the HMBC spec-
trum of 3 and comparison with that from 2. These
showed great similarity in that the presence of the same
rha-(1!2)-gal-(1!2)-glcA-(1!3) could also be discerned
in 3 at C-3. Long range correlations between H-3 and
glcA C-1, between gal C-1 and glcA H-2, and between
rha-I C-1 and gal H-2 were observed in the HMBC
spectrum. However, additional glycosylation at the gal
C-3 position was established by the appearance of
HMBC correlation between rha-II C-1 and gal H-3.
NOESY correlations of anomeric protons of each
monosaccharide were also in agreement with these
HMBC observations. Spin systems of individual mono-
saccharide units, as shown in Table 2, were identified
using selective irradiation of easily identifiable anomeric
proton signals of each monosaccharide along with irra-
diation of other non-overlapping proton signals of
individual monosaccharides in a series of selective
1D-TOCSY and HSQC-TOCSY experiments. These
Table 2
13C chemical shift assignments and corresponding 1H NMR correlation
1
also enabled complete assignment of H and 13C chemi-
a
data of the monosaccharide region of compounds 2 and 3 in pyridine-d5
cal shifts for all monosaccharide units including
identification of the multiplicity pattern of individual
protons. Anomeric configurations for the mono-
saccharide units in 3 were first inferred from the chemical
shift values of the anomeric carbons and protons. These
suggested that both the glucuronic acid unit and the
galactose unit had a b-configuration (Miyao et al., 1996).
This was also supported by the observed coupling con-
stants (J=7.2 Hz, glcA; J=7.6 Hz, gal) of the
anomeric protons of both monosaccharides. The C-5
signals of the two rhamnosyl units at ꢀ 69.3 indicated
their a-configuration (Seo et al., 1978). Hence, the
Position
3
4
ꢀC
ꢀH (J in Hz)b
ꢀC
ꢀH (J in Hz)b
GlcA
1
2
3
4
5
6
105.5
77.8
78.5
73.9
76.7
172.8
5.02 d (7.2)
4.64 dd (8.6, 7.2)
4.38 t (8.6)
105.6
77.8
78.8
73.9
76.7
172.8
4.99 d (7.2)
4.65 dd (8.2, 7.2)
4.40 t (8.2)
4.51 dd (8.6, 8.0)
4.60 d (8.0)
4.47 dd (9.0, 8.2)
4.59 d (9.0)
Gal
1
102.1
78.0
79.3
78.0
76.7
61.5
5.86 d (7.6)
101.9
78.0
79.3
78.0
76.8
61.7
5.77 d (7.6)
2
4.50 dd (9.0, 7.6)
4.26 dd (9.0, 1.6)
3.62 dd (3.4, 1.6)
4.08 td (6.2, 3.4)
4.30 dd (11, 6.2),
4.43 dd (11, 6.2)
4.54 dd (9.0, 7.6)
4.24 dd (9.0, 1.4)
3.60 dd (3.1, 1.4)
4.09 td (6.1, 3.1)
4.34 dd (12, 6.1),
4.42 dd (12, 6.1)
foregoing enabled the characterization of
3
as
3
3b,21b,22b,24-tetrahydroxyolean-12 - en - 3 - O - a - l -
rhamnopyranosyl-(1!3)-[a-l-rhamnopyranosyl-(1!2)]-
b-d-galactopyranosyl-(1!2)-b-d-glucuronopyranoside.
The negative and positive ion HR–FABMS of 4
established its molecular formula as C60H98O27 through
quasimolecular ions observed at m/z 1249.6151 [MꢀH]ꢀ
and at m/z 1273. 6242 [M+Na]+ and 1295.6067
[M+2NaꢀH]+, respectively. This, together with the 1H
and 13C NMR spectra and comparison of the same with
those of 3 suggested the presence of a third deoxyhexose
residue in addition to the triterpene aglycon and the
tetrasaccharide residue found in 3. The aglycon moiety
region in both the 1H and 13C spectra of 4 was identical
to that of 2 and differed from 3 in that a marked
downfield shift was observed for C-21. This implied
glycosylation at this position. Acid hydrolysis of the
intact saponin gave soyasapogenol A as the aglycon and
glucuronic acid, galactose and rhamnose as the mono-
saccharide residues comprising 4. However, also
observed in both the 1H and 13C spectra were five
anomeric proton and carbon signals at ꢀ 4.99, 5.50,
5.77, 6.26 and 6.38 in the 1H and at ꢀ 101.9, 102.3,
102.7, 104.7 and 105.6 in the 13C spectrum, respectively.
As in 3, HMBC correlations between glcA C-1 and H-3,
between gal C-1 and glcA H-2, between rha-I C-1 and
gal H-2, and between rha-II C-1 and gal H-3 were also
observed. On the other hand, as in 2, long-range
4
5
6
Rha-I
1
2
3
4
5
6
102.7
72.3
72.7
74.3
69.3
19.2
6.26 br s
4.70 d (3.1)
102.7
72.5
72.6
74.5
69.6
19.2
6.26 br s
4.70 d (3.2)
4.81 dd (9.5, 3.1)
4.32 dd (9.5, 9.5)
5.00 dq (9.5, 6.8)
1.80 d (6.8)
4.78 dd (9.6, 3.2)
4.34 dd (9.7, 9.6)
5.01 dq (9.7, 7.0)
1.79 d (7.0)
Rha-II
1
2
3
4
5
6
102.3
72.4
72.8
74.3
69.3
19.2
6.38
4.69 d (3.1)
102.3
72.5
72.6
74.0
69.6
19.2
6.38 br s
4.68 d (3.2)
4.79 dd (9.5, 3.1)
4.32 dd (9.5, 9.5)
5.01 dq (9.5, 6.8)
1.80 d (6.8)
4.76 dd (9.6, 3.2)
4.35 dd (9.7, 9.6)
5.02 dq (9.7, 6.9)
1.79 d (6.9)
Rha-III
1
2
3
4
5
6
104.7
72.6
72.9
74.6
69.6
18.7
5.50 br s
4.69 d (3.1)
4.61 dd (9.3, 3.1)
4.32 dd (9.4, 9.3)
4.52 dq (9.4, 6.0)
1.65 d (6.0)
a
Assignments based on 1H, 13C, DEPT, selective 1D-TOCSY and
HSQC-TOCSY experiments.
b
Multiplicities as determined by selective 1D-TOCSY.

326
G.M. Woldemichael, M. Wink / Phytochemistry 60 (2002) 323–327
correlation between rha-III C-1 and H-21 was also seen
in the HMBC spectrum. Here too, NOESY correlations
of glcA H-1 with H-3, gal H-1 with glcA H-2, rha-I H-1
with gal H-2, rha-II H-1 with gal H-3, and rha-III H-1
with H-21 were in agreement with the HMBC findings.
These assignments led to the determination of the
structure of 4 as 3b,21b,22b,24-tetrahydroxyolean-12-en-
3-O-a-l-rhamnopyranosyl-(1!3)-[a-l-rhamnopyranosyl-
(1!2)]-b-d-galactopyranosyl-(1!2)-b-d-glucuronopyr-
anosyl-21-O-a-l-rhamnopyranoside.
specimen has been kept at the Institute of Pharma-
ceutical Biology, University of Heidelberg.
4.3. Extraction and isolation
The powdered seeds of L. angustifolius (1 kg) were
ꢂ
extracted successively with petroleum ether (40–60 C)
and 70% EtOH at room temperature with constant
stirring. The EtOH extract was then concentrated to a
ꢂ
brown viscous mass under reduced pressure at 37 C.
This was suspended in H2O and partitioned with water-
saturated n-BuOH. Solvent was removed from the
n-BuOH fraction and the residue dissolved in a small
amount of MeOH, applied onto a Sephadex column
and eluted with MeOH. Fractions 7–10 were combined
and chromatographed on silica gel (EtOAc–EtOH–
H2O, 7:2:2, upper) to give 1 (30 mg) and a mixture of 2,
3 and 4. The mixture was then separated using HPLC to
yield 2 (14 mg), 3 (2 mg) and 4 (2 mg).
3. Bioactivity evaluation
The ability of 1, 2, 3 and 4 to hemolyse sheep ery-
throcytes and their activity against Candida albicans
were evaluated in vitro as described in Experimental.
The hemolysis test showed that while 1 was weakly
hemolytic (HC50 >500 mg/ml), 2, 3 and 4 did not exhi-
bit any hemolytic effect. However, in antifungal tests
against C. albicans 2, 3 and 4 showed moderate anti-
fungal activity with MICs of 25, 25 and 30 mg/ml,
respectively, while 1 was found inactive at the con-
centrations tested. The positive control used, ampho-
tericin B, was found to inhibit growth of C. albicans at
1 mg/ml.
4.4. Acid hydrolysis of 1–3
An amount of the saponin (1–5 mg) and 1 ml of 1N
HCl (dioxane–H2O, 1:1) were placed in an ampoule.
ꢂ
The ampoule was sealed and placed in an oven at 90 C
for 2 h. The ampoule was cooled, its seal broken and its
contents dried under a stream of nitrogen. Contents of
the ampoule were then suspended in H2O (3 ml), the
aqueous hydrolysate neutralized with Ag2CO3 and then
washed with EtOAc (3 ꢁ 3 ml). Removal of the solvent
after collection of the EtOAc fraction yielded the agly-
con while centrifugation followed by removal of H2O
from the aqueous portion gave monosaccharides.
Sugars were analyzed by co-chromatography with a
standard on TLC [CHCl3–MeOH–H2O (15:6:2 lower
layer)–HOAc (9:1)] and also after silylation [N-methyl-
N-(trimethylsilyl)-2,2,2-trifluoroacetamide, Sigma] using
GC–MS.
4. Experimental
4.1. General
CC: silica gel 60 (Macherey-Nagel) 0.063–0.2 mm and
Sephadex LH-20 (Pharmacia Biotech). TLC: silica gel
60 F254 (Merck). Optical rotation: Jasco P1020 digital
polarimeter. IR: Perkin–Elmer FTIR 1600 (film). GLC-
MS: Carlo Erba HRGC 4160 with OV-1 fused capillary
column (0.32 mm i.d., 0.25 mm ꢁ 30 m) and Finnigan
MAT 4500 mass spectrometer (at 45 eV). HPLC:
Knauer model 64 system with Lichrospher 100, RP-18
(5 or 10 mm), either 250 ꢁ 4 mm or 250 ꢁ 8 mm i.d.
using CH3CN – 0.1% trifluoroacetic acid in H2O linear
gradient (80 to 50% in 40 min), detection at 200 nm.
HR–FABMS: Jeol JMS 700 spectrometer with either
4.5. Compound 3
A white amorphous solid, [ꢁ]2D5 ꢀ2.9ꢂ (MeOH: c 0.83).
HR–FABMS [MꢀH]ꢀ at m/z 1103.5560 (calc. 1103.5613),
[M+Na]+ at m/z 1127.5704 (calc. 1127.5589), and
[M+2NaꢀH]+ at 1149.5493 (calc. 1149.5409). IR nmax
(film) cmꢀ1: 3358 (OH), 2958 (CH), 1649 (COOH). 13C
NMR: see Tables 1 and 2. 1H NMR NMR: see Tables 1
and 2.
PEG 100 or NBA as the matrix. H, 13C, DEPT-135,
1
DEPT-90, DQF-COSY, NOESY (mixing time 300 and
700 ms), 1D-TOCSY (for monosaccharide residues,
mixing time 60.9, 71.0 and 81.2 ms), HSQC-TOCSY
(mixing time 70 ms), HMQC and HMBC NMR spectra:
Bruker DRX 500 or DRX 600 spectrometer in pyridine-
d5 and TMS internal standard.
4.6. Compound 4
A white amorphous solid, [ꢁ]2D5 ꢀ8.4ꢂ (MeOH: c 0.83).
HR–FABMS [MꢀH]ꢀ at m/z 1249.6151 (calc.
1249.6189), [M+Na]+ at m/z 1273.6242 (calc.1273.6165),
and [M+2NaꢀH]+ at 1295.6067 (calc. 1295.5985). IR
4.2. Plant material
The seeds of L. angustifolius, available commercially,
were purchased in Germany in February 1996. Voucher
n
max (film) cmꢀ1: 3385 (OH), 2959 (CH), 1692 (COOH).

G.M. Woldemichael, M. Wink / Phytochemistry 60 (2002) 323–327
327
1
13C NMR: see Tables 1 and 2. H NMR: see Tables 1
and 2.
measurement of some of the HR–FAB spectra. One of
us (G.M.W.) gratefully acknowledges KAAD for a
doctoral fellowship while carrying out the work.
4.7. Activity against C. albicans
Compounds 1, 2, 3 and 4 were tested for antifungal
activity according to a modified NCCLS micro dilution
assay procedure. All tests were carried out using sterile
96-well plates and Saburaud dextrose broth-SDB
(Sigma). The log phase inoculum was obtained by incu-
References
Arao, T., Kinjo, J., Nohara, T., Isobe, R., 1997. Oleanane type tri-
terpene glycosides from Puerariae radix: six new saponins from
Pueraria lobata. Chem. Pharm. Bull. 45, 362–366.
Bialy, Z., Jurzysta, M., Oleszek, W., Piacente, S., Pizza, C., 1999.
Saponins in alfalfa (Medicago sativa L.) root and their structural
elucidation. J. Agric. Food Chem. 47, 3185–3192.
ꢂ
bating C. albicans (ATCC 10231) for 48 h at 30 C on
Saburaud dextrose agar plates. The compounds were
dissolved in either H2O or MeOH–H2O combination.
Solutions of each test compound were further diluted
with SDB yielding their final concentration. Each sapo-
nin was evaluated in triplicate in a dose–response format
where the final concentrations were 1, 25, 50, 100 and 500
mg/ml and where the final fungal concentration was 2 ꢁ
103 cfu. To each well containing double the final sapo-
nin dilution, an equal volume of the inoculum suspen-
sion containing double the final fungal concentration
was added. The minimum inhibitory concentration
(MIC), the lowest concentration of the test compound
resulting in the complete inhibition of growth, was
determined visually. Amphotericin B (Sigma) at 1 mg/ml
final concentration was used as a positive control.
Connolly, J.D., Hill, R.A., 1991. Triterpenoids. In: Dey, P.M., Har-
borne, J.B. (Eds.), Methods in Plant Biochemistry, Terpenoids, Vol.
7. Academic Press, London, pp. 331–368.
Hosny, M., Rosazza, J.P.N., 1999. Novel isoflavone, cinnamic acid
and triterpenoid glycosides in soybean molasses. J. Nat. Prod. 62,
853–858.
Kubo, H., Kobayashi, J., Higashiyama, K., Kamei, J., Fuji, Y.,
Ohmiya, S., 2000. The hypoglycemic effect of (7R*,9aS*)-7-phenyl-
octahydroquinolizin-2-one in mice. Biol. Pharm. Bull. 23, 1114–
1117.
Kusmenoglu, S., Erdemoglu, N., Sener, B., Mulayim, M., Ozkay-
nak, I., 1998. Investigation of fatty acids from the seeds of Lupi-
nus species by gas chromatography. J. Fac. Pharm. Gazi Univ. 15,
1–5.
Miyao, H., Sakai, Y., Takeshita, T., Ito, Y., Kinjo, J., Nohara, T.,
1996. Triterpene saponins from Abrus cantoniensis (Leguminosae)
II. Characterization of six new saponins having a branched sugar.
Chem. Pharm. Bull. 44, 1228–1231.
Mohamed, M.H., Kamel, M.S., El-Moghazy, S.A., Murakoshi, I.,
1993. The hypoglycemic effect of some alkaloids. Bull. Fac. Pharm.
(Cairo Univ.) 31, 107–111.
4.8. Hemolysis assay
One milliliter of 1% sheep erythrocyte (Oxoid) sus-
pension in phosphate buffer saline (PBS, pH 7.4) was
incubated with 100 ml of saponin dilutions (1, 0.25, 0.05,
Pereire, F.C., Ouedraogo, R., Lebrun, P., Barbosa, R.M., Cunha,
A.P., Santos, R.M., Luis, M., 2001. Insulinitropic action of white
lupin seeds (Lupinus albus L.): effects on ion fluxes and insulin
secretion from pancreatic islets. Biomed. Res. 22, 103–109.
Ruiz, R.G., Price, K.R., Rosa, M.E., Rhodes, M.J.C., Fenwick, G.R.,
1995a. Determination of saponins in lupin seeds (Lupinus angusti-
folius) using high-performance liquid chromatography: comparison
with a gas chromatographic method. J. Liquid Chromatogr. 18,
2843–2853.
ꢂ
0.01, 0.005 and 0.001 mg/ml) in PBS at 37 C for 30
min. This mixture was centrifuged (2500 g, 30 min) and
the hemoglobin content in the supernatant was mea-
sured at 490 nm using a multiwell spectrophotometer
(Virion autoReader A) by transferring 200 ml to 96-well
microtiter plates. Parallel measurements of saponin dilu-
tions without erythrocytes, erythrocytes without saponin
dilutions and buffer without both erythrocytes and
saponin dilutions were made. The results were then
compared to a positive control sample containing ery-
throcytes in distilled water. The 50% hemolysing con-
centrations (HC50) of each saponin were then inferred
from the hemoglobin absorbance vs saponin concentra-
tion curve.
Ruiz, R.G., Price, K.R., Keith, R., Rosa, M.E., Arthur, A.E., Petter-
sen, D.S., Fenwick, G.R., 1995b. The effect of cultivar and
environment on saponin content of Australian sweet lupin seed. J.
Sci. Food Agric. 69, 347–351.
Seo, S., Tomita, Y., Tori, K., Yoshimura, Y., 1978. Determination of
the absolute configuration of a secondary hydroxy group in a chiral
secondary alcohol using glycosidation shifts in carbon-13 nuclear
magnetic resonance spectroscopy. J. Am. Chem. Soc. 100, 3331–
3339.
Villarroel, M., Biolley, E., Larenas, G., Wittig, E., Diaz, V., Munoz,
Y., 1996. Chemical, sensory and clinical characterization of lupin
marmalades. Arch. Latinoam. Nutr. 46, 234–237.
Wink, M., 1993. Allelochemical properties or raison d’eˆ tre of alka-
loids. In: Cordell, G.A. (Ed.), The Alkaloids, Vol. 43. Academic
Press, San Diego, pp. 1–118.
Acknowledgements
Wink, M., 1998. Lupinus. In: Hansel, R., Keller, K., Reichling, J.,
Rimpler, H., Schneider, G. (Eds.), Hager’s Handbuch der Pharma-
zeutischen Praxis, Vol. 5. Springer, Heidelberg, pp. 104–120.
The authors are thankful to Dr. Gross, Institute of
Organic Chemistry (University of Heidelberg) for