Artefact formation during acid hydrolysis of saponins from Medicago spp.

Article abstract of DOI:10.1016/j.phytochem.2017.02.018

Artefact compounds obtained during acid hydrolysis of saponins from Medicago spp. (Fabaceae), have been monitored and evaluated by GC-FID. Their identification has been performed by GC-MS and ¹H and ¹³C NMR. Saponins with different substituents on the triterpenic pentacyclic aglycones were considered, and their hydrolysis products were detected and quantified during 10?h of time course reaction. From soyasapogenol B glycoside the well known soyasapogenols B, C, D and F were obtained together with a previously undescribed sapogenol artefact identified as 3β,22β,24-trihydroxyolean-18(19)-en and named soyasapogenol H. From a zanhic acid saponin two major artefact compounds identified as 2β,3β,16α-trihydroxyolean-13(18)-en-23,28-dioic acid and 2β,3β,16α-trihydroxyolean-28,13β-olide-23-oic acid were obtained, together with some zanhic acid. Other compounds, detected in very small amount in the reaction mixture, were also tentatively identified based on their GC-MS and UV spectra. The other most characteristic saponins in Medicago spp., hederagenin, bayogenin and medicagenic acid glycosides, under acidic condition of hydrolysis, released instead the correspondent aglycones and generated a negligible amount of artefacts. Nature of artefacts and mechanism of their formation, involving a stable tertiary carbocation, is here proposed and discussed for the first time.

Full text of DOI:10.1016/j.phytochem.2017.02.018

Phytochemistry xxx (2017) 1e12
Contents lists available at ScienceDirect
Phytochemistry
journal homepage: www.elsevier.com/locate/phytochem
Artefact formation during acid hydrolysis of saponins from Medicago
spp.
,
Aldo Tava a *, Elisa Biazzi a, Mariella Mella b, Paolo Quadrelli b, Pinarosa Avato c
a Consiglio per la Ricerca in Agricoltura e l'Analisi dell'Economia Agraria e Centro di Ricerca per le Produzioni Foraggere e Lattiero Casearie, CREA-FLC, v.le
Piacenza 29, 26900 Lodi, Italy
b
ꢀ
Dipartimento di Chimica, Universita degli Studi di Pavia, v.le Taramelli 12, 27100 Pavia, Italy
c
ꢀ
Dipartimento di Farmacia-Scienze del Farmaco, Universita degli Studi di Bari Aldo Moro, via Orabona 4, 70125 Bari, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
Artefact compounds obtained during acid hydrolysis of saponins from Medicago spp. (Fabaceae), have
been monitored and evaluated by GC-FID. Their identification has been performed by GC-MS and 1H and
13C NMR. Saponins with different substituents on the triterpenic pentacyclic aglycones were considered,
and their hydrolysis products were detected and quantified during 10 h of time course reaction. From
soyasapogenol B glycoside the well known soyasapogenols B, C, D and F were obtained together with a
Received 25 August 2016
Received in revised form
8 February 2017
Accepted 14 February 2017
Available online xxx
previously undescribed sapogenol artefact identified as 3
soyasapogenol H. From a zanhic acid saponin two major artefact compounds identified as 2
trihydroxyolean-13(18)-en-23,28-dioic acid and ,3 ,16 -trihydroxyolean-28,13 -olide-23-oic acid
b,22
b
,24-trihydroxyolean-18(19)-en and named
b
,3b,16
a-
Keywords:
Medicago spp.
Fabaceae
Triterpenic pentacyclic saponins
Acid hydrolysis
Artefact formation
Chemical structure
Soyasapogenol H
GC-MS
2
b
b
a
b
were obtained, together with some zanhic acid. Other compounds, detected in very small amount in the
reaction mixture, were also tentatively identified based on their GC-MS and UV spectra. The other most
characteristic saponins in Medicago spp., hederagenin, bayogenin and medicagenic acid glycosides, under
acidic condition of hydrolysis, released instead the correspondent aglycones and generated a negligible
amount of artefacts. Nature of artefacts and mechanism of their formation, involving a stable tertiary
carbocation, is here proposed and discussed for the first time.
© 2017 Elsevier Ltd. All rights reserved.
NMR
1. Introduction
techniques (Witkowska et al., 2008). Sapogenins should also be
considered as part of the most complex structure of saponins when
The analysis of sapogenins is one of the method used to evaluate
the saponin content in saponin-rich plants such as in Medicago spp.
(Fabaceae) and other legumes. Sapogenins are released after acid
hydrolyses of saponins, functionalised (methylated and acetylated
or silylated) and then identified by GC-MS and quantified by GC-FID
generally using an internal standard (Jurzysta and Jurzysta, 1978;
Brawn et al., 1981; Rao and Bories, 1987; Tava et al., 1993, 1999;
Tava and Pecetti, 1998; Pecetti et al., 2006). Although this method
do not allow to obtain information on the sugar moieties of the
saponins, it is a better and faster approach to quantify and
discriminate among the different sapogenin (saponin) content, as
an alternative to HPLC analyses (Nowacka and Oleszek, 1994),
capillary electrophoresis separation (Tava et al., 2000), and LC-ESI-
MS (Huhman et al., 2005; Kapusta et al., 2005) as well as MALDI
elucidating their chemical structure, and they are generally iden-
tified based on their gas-chromatographic properties, their GC-MS
fragmentation patterns (Budzikiewicz et al., 1963) and by their
NMR spectra, in which multiplicities of signals is lower compared to
the corresponding spectra of saponins. Aglycone moieties were also
considered in studies focussed on the biosynthesis of these speci-
alised metabolites (Carelli et al., 2011; Tava et al., 2011; Fukushima
et al., 2013; Moses et al., 2014; Biazzi et al., 2015).
Investigation on chemical structure of saponin/sapogenin is also
of fundamental importance for evaluation of their properties,
because differences in their chemical structure influence the
structure/activity relationships (Avato et al., 2006). Moreover, it has
been reported that during saponin/sapogenin manipulation a series
of chemical modifications can occur on both aglycone and sugar
portions (Massiot et al., 1996; Tava et al., 2003) especially in acidic
environment. In particular, during hydrolysis, some saponins can
produce by-product sapogenins that originated from the rear-
rangement of the triterpenic nucleus in presence of a strong acidic
* Corresponding author.
E-mail address: aldo.tava@crea.gov.it (A. Tava).
http://dx.doi.org/10.1016/j.phytochem.2017.02.018
0031-9422/© 2017 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Tava, A., et al., Artefact formation during acid hydrolysis of saponins from Medicago spp., Phytochemistry
(2017), http://dx.doi.org/10.1016/j.phytochem.2017.02.018

2
A. Tava et al. / Phytochemistry xxx (2017) 1e12
solution. It is known for example that saponins of soyasapogenol B
(e.g. soyasaponin I) gave several soyasapogenol artefacts named
soyasapogenols C, D and F (Jurzysta, 1984; Ireland and Dziedzic,
1986; Price et al., 1986; Mahato, 1991). These compounds have
been monitored during hydrolysis and identified (in some case only
partially) by means of spectroscopic techniques.
addition, knowledge of all the formed artefacts allows to exactly
quantify saponins/sapogenins in saponin-rich plants by using a
simple and fast GC technique. Finally, the study of the reaction
mechanism of these molecule models, under specific chemical
conditions might add new information to elucidate their biosyn-
thetic intermediates.
A different behaviour was instead observed for medicagenic
acid, hederagenin and bayogenin glycosydes, the other most
representative groups of saponins in several species of the Medi-
cago genus. These saponins are more stable under hydrolysis con-
ditions and release the corresponding sapogenins that are scarcely
affected by the acid environment (Jurzysta and Jurzysta, 1978;
Brawn et al., 1981; Rao and Bories, 1987; Massiot et al., 1988; Tava
et al., 1993).
2. Results and discussion
2.1. Hydrolysis of soyasaponin I (1)
The GC-FID and GC-MS analyses of silylated sapogenins from the
acid hydrolysis of saponin 1 (Fig. 1) showed the presence of the
sapogenols B (6), C (11), D (15) and F (13), together with the pres-
By contrast, GC investigation of hydrolysis derived products
from zanhic acid glycosides, the other dominant group of saponins
in Medicago, revealed the presence of unknown compounds in
addition to intact zanhic acid. It is reported that saponins with an
ence of another peak, further identified as 3b,22b,23-trihydrox-
yolean-18(19)-en and named soyasapogenol H (16) (Fig. 2).
Although most artefacts from saponins are known since a long
time, we felt however that some of the major compounds lacked
adequate structural determination such as complete NMR and MS
identification, and since they are still designated as unknown
compounds, a reinvestigation of these genins is needed.
hydroxyl group at 16
a position of the triterpenic nucleus as for
zanhic acid, named echinocystic and quillaic acids, produce some
artefacts in acidic conditions, that are identified as D13(18) isomers
and 28,13
b
-olides (Kubota et al., 1969).
Compounds 6, 11, 13 and 15 were separated by silica gel column
chromatography and analysed by NMR experiments. Compound 16
was not obtained in a pure form (see experimental) and used as a
mixture of about 1:1 ratio with sapogenin F (13) for NMR in-
vestigations. All these data are reported in Table 1. By comparison of
NMR data obtained by us with those reported in literature
(although in some cases only partial data are available) (Heftmann
et al., 1979; Nes et al., 1981; Kinjo et al., 1985; Baxter et al., 1990;
Mahato, 1991) and interpretation of GC-MS spectra of derivatised
sapogenins, all the chemical structures were attributed.
In the present paper five triterpene saponins structurally
different for types of aglycone and substituents were purified from
Medicago spp., including M. arabica (L.) Huds., M. arborea L. and
M. sativa L., and used in this study to investigate their stability
under acid hydrolysis, a common procedure for the determination
of the aglycone moiety. Formation of sapogenins from pure soya-
sapogenol B and pure zanhic acid saponins was followed during
10 h hydrolysis in acidic conditions. Formed hydrolysis products,
including a previously undescribed artefact sapogenins, were fully
characterized by MS and NMR techniques. Purified glycosides of
hederagenin, bayogenin and medicagenic acid were also hydro-
lysed under the same acidic conditions for comparison purposes.
The mechanism of the artefact formation, involving a stable tertiary
carbocation, is here proposed and discussed.
MS spectra of silylated compounds showed the molecular ion
[M]þ with a relative intensity of 2e8%. The trimethylsilyl de-
rivatives of soyasapogenols B (6a) and C (11a) were easily identified
by the expected retro Diels-Alder fragmentation, typical of the D12
-
unsaturated oleananes (Budzikiewicz et al., 1963), with typical ions
of m/z ¼ 306 and m/z ¼ 216, respectively. The rupture of ring C of
the pentacyclic triterpenic structure was also detectable in the MS
spectra of the trimethylsilyl derivatives of soyasapogenols D (15a)
and F (13a), proved by the presence of ions m/z ¼ 278 and 203 in
The present study aims to improve poor and incomplete data
available in the literature on saponin stability under derivatization
experimental procedures with the purpose to avoid erroneous
structural attributions following their isolation and processing. In
Fig. 1. Chemical structure of purified saponins (1e5) used in this investigation and relative sapogenins (6e10). I, soyasapogenol B; II, hederagenin; III, bayogenin, IV, medicagenic
acid; V, zanhic acid.
Please cite this article in press as: Tava, A., et al., Artefact formation during acid hydrolysis of saponins from Medicago spp., Phytochemistry
(2017), http://dx.doi.org/10.1016/j.phytochem.2017.02.018

A. Tava et al. / Phytochemistry xxx (2017) 1e12
3
Fig. 2. Proposed mechanism of artefact sapogenols formation from a soyasapogenol B saponin 1 (soyasaponin I). 6: soyasapogenol B; 11: soyasapogenol C; 13: soyasapogenol F; 15:
soyasapogenol D; 16: soyasapogenol H.
both spectra, and by the rupture of ring E giving a typical fragment
ions at m/z ¼ 99 for 15a and at m/z ¼ 157 for 13a that contains the
substituent at C-22 (eOCH3 for 15a and eOSi(CH3)3 for 13a).
Other relevant information on the chemical structures of these
compounds came from NMR experiments. Based on the presence of
seven methyl groups in all the examined sapogenins, other func-
tional groups, such as double bonds and hydroxyl substituents on
the triterpenic nucleus were evidenced in both 1H and 13C spectra.
In the 13C NMR spectrum of soyasapogenol B (6) the two sec-
ondary alcoholic groups were registered at dC 77.2 and dC 81.5 while
the primary alcoholic group at dC 65.6 and all confirmed also by
DEPT experiments. The double bond was evidenced by the two
carbon resonances at dC 123.9 and dC 145.5 in the 13C NMR spec-
trum, while the presence of a vinyl proton triplet was revealed at dH
5.25 in the 1H NMR spectrum (see Table 1). These signals are typical
Compound 16 was tentatively purified from the reaction
mixture of sapogenins by means of different chromatographic
techniques such as normal, reversed-phase and ion chromatog-
raphy (see experimental), but only a mixture of about 1:1 ratio
with soyasapogenol F (13) was obtained. This mixture was used for
GC-MS and NMR experiments. Compound 16 was well separated
under the GC conditions and its MS spectrum showed the same
MW of soyasapogenols B (6) and F (13) (C39H74Si3O3, m/z ¼ 674)
but no ions from the retro Diels-Alder fragmentation were
observed, as for soyasapogenol F (13). By comparison of NMR
spectra of the mixture of compounds 13 and 16 with those of pure
soyasapogenol F (13), signals of compound 16 could be well
extracted. In the 13C NMR spectrum the presence of 30 carbon
atoms was revealed, of which seven methyl signals (dC 15.5, dC 16.8,
dC 18.0, dC 18.4, dC 23.3, dC 30.4 and dC 32.5), two secondary alco-
holic groups (dC 77.0 and dC 81.2), one primary alcoholic group (dC
65.3) and a double bond (dC 143.0 and dC 130.2) were evidenced in
DEPT experiments. By comparison of the carbon resonances of
compound 16 with those of soyasapogenols B (6) and F (13) (see
Table 1), the same triterpenic pentacyclic nucleus can be deducted,
but the double bond signals, registered at different resonances
compared to that of soyasapogenols B (6) and F (13), let us to hy-
pothesize its different position in the molecule. DEPT experiments
clearly identified these signals as a CH and a quaternary carbon
atom. In the 1H NMR spectrum of the mixture of compounds 13
and 16, a singlet vinyl proton at dH 4.86, not present in the 1H NMR
spectrum of pure soyasapogenol F (13) was registered, and
attributed to H-19 of compound 16. Based on these findings and by
comparison of carbon resonances with data available from litera-
ture (Mahato and Kundu, 1994) the double bond in the triterpenic
of the
et al., 1990; Mahato and Kundu, 1994).
The same resonances for
D
12-unsaturated oleananes (Kojima and Ogura, 1989; Baxter
D
12(13) double bond were registered in
both 1H and 13C NMR spectra of soyasapogenol C (11), together with
signals at dC 135.0 and dC 136.8 in the 13C NMR spectrum that
correlate to the vinyl proton signals at dH 5.21 and dH 5.27 in the 1H
NMR spectrum, confirming the presence of the additional double
bond D
21(22) in the molecule. Only one primary alcoholic group was
found at dC 65.1 while the secondary alcoholic group resonated at dC
80.9 (Nes et al., 1981).
The carbon resonances of D
13(18) double bond of soyasapogenol
F (13) were registered at dC 133.8 and dC 138.6, while the allyl sig-
nals, H-12 and H-19, were well evidenced at dH 2.71 and dH 1.85, and
dH 2.34 and dH 1.70 in both the 1H NMR and 2D NMR spectra (see
Table 1). The three alcoholic carbon resonances of 13 were found at
dC 65.5, dC 78.8 and dC 81.4.
nucleus of 16 was attributed to
identified as ,22 ,24-trihydroxyolean-18(19)-en and named
soyasapogenol H (16).
D
18(19) position. This compound was
The same D13(18) double bond carbon resonances were regis-
tered in the NMR spectrum of soyasapogenol D (15) (see Table 1).
The two alcoholic functions in the molecule were evidenced at dC
65.7 and dC 81.6, while the presence of the methoxy group was
confirmed by the presence of the corresponding signals at dH 3.34
and dC 58.7 in the 1H and 13C NMR spectra, respectively.
3
b
b
The mechanism of formation of soyasaponin I (1) artefacts is
reported in Fig. 2, while the quantitative evaluation of conversion of
soyasaponin I (1) into sapogenins during 10 h of acid hydrolysis is
reported in Fig. 3.
Please cite this article in press as: Tava, A., et al., Artefact formation during acid hydrolysis of saponins from Medicago spp., Phytochemistry
(2017), http://dx.doi.org/10.1016/j.phytochem.2017.02.018

4
A. Tava et al. / Phytochemistry xxx (2017) 1e12
Table 1
NMR data (
d
, CD3OD/CDCl3 2:1) of soyasapogenol B (6), soyasapogenol C (11), soyasapogenol F (13), soyasapogenol D (15) and soyasapogenol H (16).
Position 1H
13C
6
11
13
15
16
6
11
13
15
16
1
2
3
1.05 and 1.64, 2Ha 1.10 and 1.71, 2Ha
1.21 and 1.80, 2Ha 1.22 and 1.85, 2Ha
1.03 and 1.29, 2Ha
1.17 and 1.70, 2Ha
3.37, 1H, dd (11.2, 4.5) 3.39, 1H, dd (11.1, 4.5) 3.38, 1H, dd
(11.0, 4.3)
1.00 and 1.25, 2Ha
1.12 and 1.68, 1Ha
1.00 and 1.63, 2Ha 40.0
1.18 and 1.74, 2Ha 28.6
39.1
27.9
80.9
40.1
27.5
81.4
40.2
27.6
81.6
39.8
28.3
81.2
3.39, 1H, dd
(11.8, 4.1)
e
3.39, 1H, dd
(11.8, 4.2)
e
81.5
4
5
6
7
8
9
10
11
12
e
e
e
43.8
57.6
42.9
56.6
19.2
34.0
40.3
48.5
37.4
24.4
43.9
57.5
20.2
34.8
41.8
52.4
38.5
23.3
43.9
57.5
20.2
35.0
41.8
52.4
38.5
23.4
27.2
43.5
57.1
19.5
35.4
40.6
51.9
38.0
22.4
26.7
0.97, 1Ha
0.95, 1Ha
0.87, 1Ha
0.89, 1H, bd
0.86, 1H, bd
1.36 and 1.65, 2Ha 1.39 and 1.63, 2Ha
1.24 and 1.63, 2Ha
1.22 and 1.65, 2Ha
1.28 and 1.62, 2Ha 20.1
1.43 and 1.73, 2Ha 34.8
1.42 and 1.58, 2Ha 1.44 and 1.61, 2Ha
1.41 and 1.67, 2Ha
1.35 and 1.82, 2Ha
e
e
e
e
e
41.1
49.5
38.2
1.58, 1Ha
1.61, 1Ha
1.52, 1H, bd
e
1.48, 1H, db
e
1.50, 1H, db
e
e
e
1.57 and 1.88, 2Ha 1.51 and 1.83, 2Ha
1.22 and 1.53, 2Ha
1.20 and 1.48, 2Ha
1.21 and 1.50, 2Ha 25.1
5.24, 1H, t (3.3)
5.30, 1H, t (3.3)
1.88, 1Ha and 2.71,
1.79, 1Ha and 2.67,
1.25 and 1.55, 2Ha 123.9 123.2 27.1
1H, dm (15.3)
e
1H, dm (15.1)
e
13
14
15
16
17
18
e
e
e
e
1.65, 1Ha
145.5 145.1 138.6 138.6 39.5
e
e
43.6
27.2
43.0
26.3
30.6
35.7
46.8
45.7
28.8
36.7
42.6
45.8
28.8
36.8
42.7
43.5
28.4
36.0
41.7
1.04 and 1.76, 2Ha 1.10 and 1.78, 2Ha
1.65e1.80, 2Ha
1.67e1.86, 2Ha
1.65e1.85, 2Ha
1.29 and 1.76, 2Ha 1.25 and 1.76, 2Ha
e
1.43e1.52, 2Ha
e
1.40e1.49, 2Ha
e
1.28 and 1.75, 2Ha 30.2
e
e
38.7
47.0
2.05, 1H, dd
(14.0, 4.0)
2.12, 1H, dd (14.0, 4.0)
e
e
e
133.8 133.6 143.0
19
0.96 and 1.76, 2Ha 1.04 and 1.81, 2Ha
1.71, 1Ha and 2.34,
1H, dd (12.0, 2.1)
e
1.38 and 1.53, 2Ha
3.28, 1H, dd (12.0, 4.5) 2.85, 1H, dd (12.0, 4.5) 3.59, 1H, dd
(12.0, 3.9)
1.22, 3H, s
3.39 and 4.10,
2H, d (11.4)
0.89, 3H, s
0.90, 3H, s
1.22, 3H, s
0.99, 3H, s
0.78, 3H, s
0.99, 3H, s
e
1.69, 1Ha and 2.31,
1H, dd (12.0, 2.1)
e
1.33 and 1.59, 2Ha
4.86, 1H, s
47.8
31.7
47.1
33.8
136.8 44.9
135.0 78.8
39.3
33.4
39.6
130.2
20
21
22
e
e
e
33.3
40.7
89.6
34.4
41.8
77.0
1.34 and 1.45, 2Ha 5.27, 1H, d (9.9)
1.38 and 1.55, 2Ha 42.5
3.37, 1H, dd
(12.0, 4.5)
1.19, 3H, s
3.38 and 4.10,
2H, d (11.9)
0.95, 3H, s
0.97, 3H, s
1.12, 3H, s
1.01, 3H, s
0.91, 3H, s
0.83, 3H, s
e
5.20, 1H, d (9.9)
77.2
23
24
1.19, 3H, s
3.39 and 4.19,
2H, d (11.9)
0.95, 3H, s
0.97, 3H, s
1.12, 3H, s
1.01, 3H, s
0.91, 3H, s
0.83, 3H, s
e
1.21, 3H, s
3.34 and 4.31,
2H, d (11.1)
0.85, 3H, s
0.86, 3H, s
1.18, 3H, s
0.98, 3H, s
0.76, 3H, s
1.00, 3H, s
3.34, 3H, s
1.24, 3H, s
3.32 and 4.16,
2H, d (11.1)
0.87, 3H, s
0.88, 3H, s
0.77, 3H, s
0.99, 3H, s
1.04, 3H, s
1.09, 3H, s
e
23.5
65.6
25.2
65.1
23.5
65.5
23.7
65.7
23.3
65.3
25
26
27
28
29
30
OCH3
16.9
17.8
25.7
29.3
32.9
20.7
e
16.8
17.5
26.6
28.5
31.9
23.1
e
18.5
17.6
22.0
17.3
25.8
33.1
e
18.7
18.3
22.4
17.8
26.0
33.4
58.7
18.4
18.0
15.5
16.8
32.5
30.4
e
The underlined term indicates the carbon atom to which the signal belongs in a functional group. 1H NMR assignments were established by HSQC, DQF-COSY and TOCSY
experiments. J values (in hertz) are given in parentheses.
a
Multiplicities not assigned due to overlapped signals.
the amount of the single sapogenin was instead observed during all
the examined period of hydrolysis. Soyasapogenol B (6) showed an
increase from 1 to 2 h and then a decrease till to 10 h hydrolysis,
while a constant increase was in general observed for all the other
identified artefacts (Fig. 3). These data suggest that the so formed
aglycone 6 was than successively transformed into artefact sap-
ogenols involving a formation of a stable tertiary carbocation (12)
that, after a proton elimination, can generate sapogenins 13 and 16.
As reported by Mahato (1991), assuming all-chair conformation of 6
having D/E rings cis-fused, the 22
strong steric interactions with the 20
that of the 17-methyl group. Migration of the double bond from the
12:13 to 13:18 position transforms the 22 -hydroxyl from axial to
b
-axial hydroxyl experiences
b
-axial methyl, in addition to
b
equatorial orientation, thereby lowering the 1,3-diaxial interaction.
The same finding can be applied for the double bond migration
from the 12:13 to 18:19 position in compound 16. Substitution of
Fig. 3. Conversion of soyasaponin I 1 into sapogenins. The graph shows the mmol % of
the obtained compounds as a function of time of hydrolysis. 6: soyasapogenol B; 11:
soyasapogenol C; 13: soyasapogenol F; 15: soyasapogenol D; 16: soyasapogenol H.
the 22b-hydroxyl group in 13 with a methoxy group to give com-
pound 15, may then occur as shown in Fig. 2 via protonation and
elimination of the protonated group through participation of the
13:18 double bond which is favourably disposed to form the in-
termediate homoallylic carbocation 14. A successive attack by a
molecule of methanol on 14 generates a methoxy derivative
15 with retention of configuration. The presence of the so-called
Winstein type homoallylic carbocation has been previously
reported for other type of polycyclic compounds (Aneja et al.,
The total sapogenin content, espressed as
mmol% (Fig. 3),
increased during the first 2 h of hydrolysis and then remained
more or less constant for the rest of the reaction time, reaching
the maximum values after 8 h of hydrolysis, with a yield of
93.6 4.5 mmol% of the total sapogenins. By contrary, a variation in
Please cite this article in press as: Tava, A., et al., Artefact formation during acid hydrolysis of saponins from Medicago spp., Phytochemistry
(2017), http://dx.doi.org/10.1016/j.phytochem.2017.02.018

A. Tava et al. / Phytochemistry xxx (2017) 1e12
5
1975; Cadenas et al., 2005; Ding et al., 2011). The introduction in the
triterpenic nucleus of a RO- group was also confirmed by treatment
of saponin 1 with CD3OD and EtOH (see experimental) and com-
pounds 15b and 15c were identified based on their molecular
weight (C37H65D3Si2O3 m/z ¼ 619 and C38H70Si2O3 m/z ¼ 630,
respectively) and by the presence of ions at m/z ¼ 102 for 15b and
m/z ¼ 113 for 15c, originated from the rupture of ring E that con-
tains the above mentioned substituent at C-22 (eOCD3 for 15b and
eOCH2CH3 for 15c, respectively).
zanhic acid 10a. The presence of 30 carbon atoms was revealed in
the 13C NMR spectrum of compound 18, of which six methyl signals
at dC 13.0, dC 18.6, dC 19.6, dC 25.3, dC 26.7 and dC 33.2, three sec-
ondary alcoholic groups at dC 76.4, dC 72.1 and dC 71.9, and a double
bond (dC 137.5 and dC 128.1) were evidenced in DEPT experiments.
Two carboxylic group signals at dC 181.9 and at dC 180.2, were also
registered. By comparison of the carbon resonances of compound
18 with those of zanhic acid 10 (Table 2), the same triterpenic
pentacyclic nucleus could be inferred. Only the double bond sig-
nals, upfield shifted compared to the corresponding zanhic acid
signals, let us to deduce its different position in the molecule. In the
1H NMR spectra of 18 no vinyl protons (H-12 in zanhic acid at dH
5.33) were registered, and the allyl proton H-18 (at dH 3.04 in zanhic
acid) well evidenced in the 1H NMR of sapogenins possessing a
12e13 double bond, was absent. Two allyl signals, H-12 and H-19,
were well evidenced at dH 2.71 and dH 1.96, and dH 2.43 and dH 1.78
in both the 1H NMR and 2D NMR spectra (Table 2), indicating the
13:18 position of the double bond in the triterpenic nucleus. This
2.2. Hydrolysis of zanhic acid saponin 5
To investigate the artefacts obtained from zanhic acid saponins,
compound 5 was taken as representative of zanhic acid glycosides
and subjected to acid hydrolysis. The obtained aglycones (10,18 and
19) were purified by silica gel column chromatography and pre-
parative reversed-phase HPLC. Compounds 10 (zanhic acid), 18 and
19, were obtained in a pure grade together with several other by-
products, including mono and dimethyl esters of the triterpenic
dicarboxylic acid. Their structure elucidation was performed by
combining NMR (Table 2), GC-MS and ESI-MS-MS data.
compound was identified as 2b,3b,16a-trihydroxyolean-13(18)-en-
23,28-dioic acid (18).
From ESI-MS/MS experiments compound 19 showed a molec-
ular weight of m/z ¼ 532, corresponding to a molecular formula of
Compound 10, was easily identified as zanhic acid, 2b,3b,16a-
trihydroxyolean-12-en-23,28-dioic acid, on the basis of its MS
fragmentation pattern, its NMR data (Table 2), and by comparison
with data from the literature (Bialy et al., 1999; Kapusta et al., 2005;
Tava et al., 2005).
Compound 18 showed the same molecular weight of zanhic
acid, as deducted from the ESI-MS-MS and GC-MS spectra, but no
ions from the retro Diels-Alder fragmentation were observed for
the silyl derivative 18a, compared to the same silyl derivative of
C31H48O7, while from GC-MS of its derivative 19a, a molecular ion of
m/z ¼ 748 was revealed corresponding to a molecular formula of
C
40H72O7Si3. As for compound 18a no ions from retro Diels-Alder
fragmentation were observed. In the 13C NMR spectrum the two
carbonyl resonances at dC 180.2 and dC 180.7 were registered
together with three secondary alcoholic groups at dC 76.3, dC 71.9
and dC 70.2, that correlate with the signals at dH 3.96, dH 4.11 and dH
3.65 respectively, in the 1H NMR spectrum. No vinyl signals were
Table 2
NMR data (
d
, CD3OD) of zanhic acid (10), 2
b,3
b
,16
a-trihydroxyolean-13(18)-en-23,28-dioic acid (18) and methyl ester of 2
b,3
b
,16
a-trihydroxyolean-17 / 13-lactone-23-oic
acid (19).
Position
1H
10
13C
18
19
10
18
19
1
2
3
4
5
6
7
8
1.37, 1Ha and 2.23, 1H, dd (14.4, 3.3)
4.11, 1H, bq (3.3)
4.01, 1H, d (3.3)
1.28, 1Ha and 2.25, 1H, dd (14.4, 3.6)
4.11, 1H, bq (3.6)
4.01, 1H, d (3.6)
1.24, 1Hs and 2.23, 1H, dd (14.5, 3.4)
45.9
73.3
76.7
54.1
53.1
22.3
34.2
41.3
49.2
37.7
24.9
123.6
145.4
43.1
36.4
75.6
50.5
42.4
48.0
31.7
36.9
33.1
182.3
13.4
17.5
18.0
27.6
181.5
33.7
25.2
e
46.1
71.9
76.4
55.0
53.3
22.3
35.4
43.8
53.0
37.8
23.7
26.9
137.5
46.5
36.9
72.1
54.2
128.1
43.1
34.9
37.5
27.1
181.9
13.0
18.6
19.6
26.7
180.2
33.2
25.3
e
45.8
71.9
76.3
54.8
52.7
21.6
35.3
42.9
50.8
37.4
23.8
19.5
93.0
45.0
35.1
70.2
54.5
39.7
37.1
30.7
36.4
28.5
180.7
12.8
18.1
18.3
23.3
180.2
33.6
21.8
52.8
4.11, 1H, bq (3.4)
3.96, 1H, d (3.4)
e
e
e
1.58, 1Ha
1.55, 1Ha
1.50, 1Ha
1.20 and 1.67, 2Ha
1.18 and 1.60, 2Ha
1.52 and 1.67, 2Ha
1.25 and 1.43, 2Ha
e
1.32 and 1.57, 2Ha
1.40 and 1.50, 2Ha
e
e
9
1.67, 1Ha
1.57, 1Ha
1.33, 1Ha
e
1.65 and 1.92, 2Ha
1.54e1.65, 2Ha
e
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
COOCH3
e
e
1.95e2.00, 2Ha
1.35 and 1.61, 2Ha
5.33, 1H, bt (3.2)
e
1.96, 1Ha and 2.71, 1H, dm (14.1)
e
e
e
e
1.35 and 1.83, 2Ha
4.48, 1H, bq (4.5)
e
1.32 and 1.93, 2Ha
1.60 and 1.91, 2Ha
3.65, 1H, bq (4.4)
e
4.20, 1H, dd (12.0, 3.9)
e
3.04, 1H, dd (14.0, 3.5)
1.35 and 2.31, 2Ha
e
1.16 and 1.91, 2Ha
1.71 and 1.92, 2Ha
e
e
2.87, 1H, dd (13.0, 4.1)
0.92 and 1.35, 2Ha
e
1.07 and 1.27, 2Ha
1.70 and 1.84, 2Ha
e
1.78, 1Ha and 2.45, 1H, dd (13.5, 1.8)
e
1.30 and 1.54, 2Ha
1.16 and 1.61, 2Ha
e
1.38, 3H, s
1.35, 3H, s
0.83, 3H, s
1.41, 3H, s
e
1.32, 3H, s
1.24, 3H, s
1.01, 3H, s
1.29, 3H, s
e
1.33, 3H, s
1.23, 3H, s
1.10, 3H, s
0.87, 3H, s
e
0.91, 3H, s
1.00, 3H, s
e
0.94, 3H, s
0.80, 3H, s
e
0.89, 3H, s
1.39, 3H, s
3.71, 3H, s
The underlined term indicates the carbon atom to which the signal belongs in a functional group. 1H NMR assignments were established by HSQC, DQF-COSY and TOCSY
experiments. J values (in hertz) are given in parentheses.
a
Multiplicities not assigned due to overlapped signals.
Please cite this article in press as: Tava, A., et al., Artefact formation during acid hydrolysis of saponins from Medicago spp., Phytochemistry
(2017), http://dx.doi.org/10.1016/j.phytochem.2017.02.018

6
A. Tava et al. / Phytochemistry xxx (2017) 1e12
evidenced in both spectra, while six methyl (Table 2) and one
methoxy resonances (dC 52.8) were registered and confirmed in the
DEPT experiment. Moreover, a signal of a totally substituted
oxygenated carbon (dC 93.0) was evidenced, this last being char-
acteristic of a lactone group between C-28 and C-13 (Marx Young
et al., 1997; Martinez et al., 2015). This compound was then iden-
tified as
2b,3b,16a-trihydroxyolean-28,13b-olide-23-oic acid
methyl ester (19).
A plausible mechanism of formation of the above artefacts from
saponin 5 is reported in Fig. 4, while the quantitative evaluation of
its conversion into sapogenins during the 10 h of acid hydrolysis is
reported in Fig. 5.
The total sapogenin content, espressed as mmol% (Fig. 5), slowly
increased during the hydrolysis time reaching a maximum at
8e10 h. A variation of the single sapogenin amount was also
observed during all the examined period of hydrolysis, with zanhic
acid (10) showing an increase from 1 to 2 h and then a decrease till
to 10 h hydrolysis. A constant increase in concentration was
observed for compounds 18 and 19 (Fig. 5). These data revealed
that, as for soyasapogenol B (6), the so formed aglycone (10) was
successively transformed into artefact sapogenols involving a stable
tertiary carbocation (17). By contrast, for zanhic acid no methoxy
derivative was detected as for soyasapogenol B, and the formation
of the isomerized compound 18 and the lactone 19 clearly indicates
Fig. 5. Conversion of zanhic acid saponin 5 into sapogenins. The graph shows the
% of the obtained compounds as a function of time of hydrolysis. 10: zanhic acid; 18:
,3 ,16 -trihydroxyolean-13(18)-en-23,28-dioic acid; 19: 2 ,3 ,16 -trihydroxyolean-
28,13 -olide-23-oic acid.
mmol
2b
b
a
b
b
a
b
2.3. Hydrolysis of saponins 2, 3 and 4
the inability of the 16a-hydroxyl group to act as good leaving group,
presumably because of the lower stability of the corresponding
secondary carbocation adjacent to the carboxylic group at C28
position. Moreover, the different geometrical features can result
inapt to be stabilized through homoallylic interactions.
Hydrolysis of saponins from hederagenin (2), bayogenin (3) and
medicagenic acid (4) (Fig. 1) performed in the same acidic condi-
tions, showed the presence of one compound clearly attributed to
the corresponding aglycone (7, 8 and 9, respectively) that, during
the 10 h of hydrolysis increased in yield reaching the maximum at
7e8 h when all the saponin was completely hydrolysed (data not
showed). As for the other group of saponins, the total amounts of
aglycones slowly decreased after this time, likely due to decom-
position (Tava et al., 1993). During the hydrolysis reaction, saponins
possessing a carboxylic group in the triterpenic nucleus could un-
dergo a transesterification reaction and the amount of formed
methyl esters could be evaluated by GC-MS after a direct silylation
of the reaction mixtures (see experimental). Since discrete amounts
of methyl esters are formed from saponins during their acid
treatment (see Table 3), it is required to accomplish a complete
methylation of sapogenin mixtures before sylanization or acetyla-
tion for GC analysis to avoid their incomplete derivatization. Other
minor compounds were detected and tentatively identified based
on their GC-MS spectra (Budzikiewicz et al., 1963), and the corre-
However, other minor byproducts (21 and 22) were formed in
the reaction mixture after 30 h of hydrolysis, originated from the
elimination of the 16-hydroxyl group of zanhic acid (10). These
compounds, detected in very low amount (less than 1% of the total
mixture) were tentatively identified based on their GC-MS and UV
spectra. An hypothesis of their formation is reported in Fig. 6. The
protonation of the 16-hydroxyl group of zanhic acid (10) and
elimination of the protonated group originate the intermediate
carbocation 20 that, after a proton elimination, gave compound 21,
with a reaction mechanism very similar to that of formation of
soyasapogenol C (11). Moreover, the presence of the carboxylic
group likely could promote a different rearrangement that, after
loss of CO2, leads to the formation of conjugated 12:13, 17:18 diene
(22), well identified based on its UV absorbance (see experimental).
The presence of a conjugated diene was previously detected from
the reaction products of acidic treatment of other 16-hydroxy
pentacyclic compounds (Kubota et al., 1969).
sponding
olides (compounds 25, 27 and 29) were found. These results are
D b
13(18) isomers (compounds 24, 26 and 28) and the 28,13
Fig. 4. Proposed mechanism of formation of compounds 18 and 19 from zanhic acid saponin 5.
Please cite this article in press as: Tava, A., et al., Artefact formation during acid hydrolysis of saponins from Medicago spp., Phytochemistry
(2017), http://dx.doi.org/10.1016/j.phytochem.2017.02.018

A. Tava et al. / Phytochemistry xxx (2017) 1e12
7
Fig. 6. Proposed mechanism of formation of compounds 21 and 22 from zanhic acid 10.
Table 3
The presence of an eOH group in 16 or 22 position (
g
position
Percentage composition of artefacts obtained from the different compounds after 8
and 30 h of acid hydrolyses.
to the double bond) of the triterpenic aglycone promotes this
transposition leading to a lower steric interaction with the
methyl substituents. In case of soyasapogenol B other types of
rearrangement can be observed as the formation of the related
isomer D18(19) (soyasapogenol H) or the introduction in the
molecule of nucleophiles (soyasapogenol D). In addition, the
presence of the carboxylic group adjacent to the carbocation, as
for zanhic acid, induce instead the formation of the correspond-
8 h
30 h
Hederagenin
Bayogenin
Methyl estera
1e3
2e5
5e9
8e12
3e5
70e76
4e7
7e9
D13(18) (24)
28,13
b-olide (25)
e
Untransformed compound (7)
87e91
1e4
2e3
Methyl estera
D13(18) (27)
ing
g-lactone.
28,13
b
-olide (28)
e
3e5
Untransformed compound (8)
Mono- and di- methyl estersa
D13(18) (30)
88e92
8e10
2e3
70e73
20e25
9e11
2e4
65e72
15e22
73e78
14e20
2e5
Knowledge acquired through our study on formation of saponin
Medicagenic acid
Zanhic acid
artefacts under acidic hydrolysis and structural characterization of
by-products formed should then be regarded of importance to
discriminate among the presence of artefact compounds in saponin
containing plants and avoid mistaken structural determinations.
In addition, the knowledge of nature of artefacts obtained from
acid hydrolysis of a particular saponin is fundamental for the exact
quantification of that saponin. The sum of all the amounts from the
GC peaks gave the exact content of that particular compound from
which all artefacts originates. This helps to obtain an appropriate
quantitative determination of saponins/sapogenins in plant
matrices, especially for the accurate determination of the most
common soyasaponin I in legumes which otherwise will be
underestimated.
28,13
b-olide (31)
e
Untransformed compound (9)
Mono and di- methyl estersa
D13(18) (18)
79e87
7e10
68e71
6e11
11e18
28,13
b-olide (19)
Untransformed compound (10)
Including D13(18) and 28,13
-olide derivatives.
b
a
summarized in Table 3. For all the examined compounds a similar
reaction mechanism involving a stable tertiary carbocation (23) can
be presumed as outlined in Fig. 7.
Finally, the elucidation of the mechanism of reaction of the more
unstable saponin structural types possibly allows to understand the
chemical behaviour of similar natural products (eg. triterpenes
3. Concluding remarks
with an eOH group in
g position to the double bond, such as
Our study on the stability of Medicago saponins under common
derivatization procedures showed that some molecular types, in
particular soyasapogenols and zanhic acid, forms artefacts under
acidic hydrolysis. Thus, availability of triterpenic saponins struc-
turally different for types of aglycone and substituents allowed to
investigate the reaction mechanisms involved in the formation of
these by-products and characterize their structure based on the
reactant saponin.
caulophyllogenin, quillaic and echinocystic acids) which can un-
dergo similar degradation. The characterization of these artefacts
can add new evidences to understand triterpene saponins biosyn-
thetic steps as well as to investigate the production of new mole-
cules (including artefacts) with potential industrial interest.
4. Experimental
Results obtained indicated that artefacts production involves an
intermediate tertiary carbocation, formed from the aglycone, which
leads double bond transposition from 12:13 to 13:18 position in the
triterpenic structure, affording these by-products in different
amount depending on the chemical nature of the sapogenin under
reaction.
4.1. General experimental procedures
All pure compounds and fractions from the chromatographic
steps were analysed by TLC, GC-FID, GC-MS, HPLC and NMR
methods. Merck silica gel 60H were used for TLC and sapogenins
Please cite this article in press as: Tava, A., et al., Artefact formation during acid hydrolysis of saponins from Medicago spp., Phytochemistry
(2017), http://dx.doi.org/10.1016/j.phytochem.2017.02.018

8
A. Tava et al. / Phytochemistry xxx (2017) 1e12
Fig. 7. Proposed mechanism of formation of D13(18) isomers (24, 26 and 28) and 28,13
acid (4).
b-olides (25, 27 and 29) from saponins of hederagenin (2), bayogenin (3), and medicagenic
were eluted with petroleum ether/CHCl3/AcOH (7:2:1) or benzene/
MeOH (9:1) and spots visualized by spraying with MeOH/acetic
anhydride/sulfuric acid (10:1:1 v/v) followed by heating at 120 ꢀC.
GC-FID and GC-MS were performed on sapogenins as their
methyl-silyl derivatives as described in Tava et al., 2005. GC-FID
analyses were carried out using a Perkin-Elmer model 8500 GC
of 70 eV. Transfer line temperature was 300 ꢀC, and the carrier gas
was He at 1.2 mL/min.
Sapogenins were also submitted to HPLC analyses using a Perkin
Elmer chromatograph equipped with a LC250 binary pump and
DAD 235 detector. Separation was performed on a Discovery HS-
C18 column (Supelco, 250 ꢁ 4.6 mm, 5
mm) with the following
equipped with a 30 m ꢁ 0.32 mm i.d., 0.25
m
m, DB-5 capillary
mobile phase: solvent A: CH3CN, 0.05% CF3COOH; solvent B: H2O,
1% MeOH, 0.05% CF3COOH. Chromatographic runs were carried out
under gradient elution from 50% (1 min isocratic condition) to 100%
column. Injector and detector temperatures were set at 350 ꢀC, and
the oven temperature program was as follows: 90 ꢀC for 5 min,
increased at 20 ꢀC/min to 250 ꢀC for 1 min and then increased at
of solvent A in 20 min and remaining at 100% of A for 30 min. 20 ml
4
ꢀC/min to 350 ꢀC for 15 min. Samples (1
ml) were injected in the
of methanolic solutions (1 mg/ml) of all samples were injected.
Sapogenins were eluted at 1.0 ml/min and detected by UV moni-
toring at 215 nm. UV spectra were collected from 190 to 350 nm.
ESI-MS-MS analyses were performed on a 1100 Series Agilent
LC-MSD Trap-System VL. An Agilent Chemstation (LC-MSD trap-
Software 4.1) was used for acquisition and processing of the data.
All the analyses were carried out using a ESI ion source type in the
splitless mode. He was the carrier gas with a head pressure of 12.2
psi. GC-MS analyses were carried out using a Perkin-Elmer Clarus
500 GC equipped with a MS detector and a 30 m ꢁ 0.25 mm i.d.,
0.25
mm, Elite-5MS capillary column using the same chromato-
graphic conditions as for GC-FID. Mass spectra were acquired over
the 50e850 amu range at 1 scan/s with an ionizing electron energy
Please cite this article in press as: Tava, A., et al., Artefact formation during acid hydrolysis of saponins from Medicago spp., Phytochemistry
(2017), http://dx.doi.org/10.1016/j.phytochem.2017.02.018

A. Tava et al. / Phytochemistry xxx (2017) 1e12
9
negative mode with the following settings: capillary voltage,
4.4. Artefact monitoring during acid hydrolysis of saponins 1 and 5
4000 V; nebulizer gas (N2) 15 psi; drying gas (N2); heated at 350 ꢀC
and introduced at a flow of 5 l/min. Full scan spectra were acquired
over the range of 100e2200 m/z with a scan time of 13,000 m/z/sec.
Automated MS-MS was performed by isolating the base peak
(molecular ions) using an isolation width of 4.0 m/z, fragmentation
amplitude of 1.0 V, threshold set at 100 and ion charge control on,
with max acquire time set at 300 ms. Samples were dissolved in
MeOH:H2O (9:1) at the concentration of 20e30 ppm and injected
Saponin 1 (10.4 mg, C48H78O18, Mr 942, 10.6
(10.7 mg, C52H82O24, Mr 1090, 9.2 mol) were separately treated with
30 mL of 2 N HCl in MeOH:H2O 1:1 under reflux. Every 2 h, 5 ml of
solution, corresponding to 1.73 mg (1.84 mol) of saponin 1 and
1.78 mg (1.64 mol) of saponin 5, were sampled aftercooling, 0.71 mg
(1.61 mol) of uvaol was added as internal standard and thereafter
aglycones were extracted by using ethyl acetate (3 ꢁ 2 ml). The sol-
vent was evaporated to dryness and the obtained sapogenins were
methylated with CH2N2 and then silylated (0.2 ml of pyridin/hex-
amethyldisilazane/chlorotrimethylsilane 2:1:1 and heated at 70 ꢀC
for 10 min) before GC injections. Hydrolysis was performed in trip-
licateandsolutionsusedseparatelyforGC-FIDandGC-MSevaluation.
mmol) and saponin 5
m
m
m
m
by direct infusion at a flow rate of 10 ml/min (KDScientific Syringe
Pump).
1H and 13C NMR were measured on a Bruker AV-300 spec-
trometer at the operating frequencies of 300.13 and 75.13 MHz,
respectively. Sapogenins were examined as solutions in CD3OD/
CDCl3 2:1 (5e10 mg/0.5 ml) in 5 mm tubes at 25 ꢀC. TMS was used
as internal reference. 2D NMR experiments (H,H DQF-COSY; H,H
TOCSY; H,H NOESY; H,H ROESY; H,C HSQC; H,C HMBC) were carried
out on all compounds using the phase sensitive method. Based on
2D NMR analyses, assignments of 1H and 13 C signals were obtained.
Melting points were determined using a Buchi apparatus and
are incorrect. Elemental analyses were carried out on a Carlo Erba
instruments. Optical rotations were measured on a Perkin-Elmer
241 polarimeter at 25 ꢀC.
4.5. Hydrolysis of saponin 1 and purification of compounds 6, 11,
13, 15 and 16
Saponin 1 (500 mg, 530.8 mmol) was treated with 300 ml of 2 N
HClin MeOH:H2O 1:1 underrefluxfor8 h. MeOH wasremoved under
reduced pressure and aglycones were extracted by using ethyl ace-
tate (3 ꢁ 100 ml). The organic solution was treated with anhydrous
Na2SO4 and the solvent removed under reduced pressure to obtain
225.8 mg (493.0
crude sapogenin mixture. The sapogenin mixture was submitted to a
400 ꢁ 55 mm, 40e60 m silica gel column (Merck). Fractions were
mmol as soyasapogenol B equivalent, 92.9% yield) of
4.2. Extraction and purification of saponins 1e5
m
M. arabica (L.) Huds., M. arborea L. and M. sativa L. (Fabaceae)
were grown at CREA-FLC, Lodi (45ꢀ 190N, 9ꢀ 300 E, 81 m elevation), a
location characterized by rather favourable, sub-continental cli-
matic conditions (802 mm long-term average annual rainfall). Leaf
sampling was carried out at flowering stage for all species, oven
dried at 50 ꢀC, powdered and used for saponin extraction. Pure
saponins 1e5 were obtained from the plant material following
previously reported procedures (Bialy et al., 1999; Tava et al., 2005,
2009). Their purity and identity were evaluated by TLC, HPLC, NMR
and ESI-MS-MS analyses (Tava et al., 2005, 2009). These saponins
eluted with CHCl3 and checked by TLC developed with petroleum
ether/CHCl3/acetic acid (7:2:1 v/v) and toluene/MeOH (85:15 v/v),
visualisingthespotsbysprayingwith MeOH/acetic anhydride/H2SO4
(10:1:1 v/v) followed by heating at 120 ꢀC. Fractions containing the
same compounds were combined and 115.3 mg of a mixture of sa-
pogenins 11 and 15 and 108.7 mg of a mixture of compound 6,13 and
16 were obtained. The first fraction was further fractionated using a
silica gelcolumnelutingwith hexane/Et2O (97:3 v/v)toobtain8.7mg
of pure compound 11, (soyasapogenol C) and 67.3 mg of pure com-
pound 15 (soyasapogenol D). The second fraction was further frac-
tionated using a silica gel column eluting with toluene/MeOH (95:5
v/v) to obtain 32.1 mg of pure compound 13, (soyasapogenol F),
5.6 mg of pure compound 6 (soyasapogenol B) and 28.7 mg of
mixture of compounds 13 and 16 in the ratio of above 1:1. Com-
pounds 13 and 16 were tentatively separated using different chro-
matographic techniques such as ion chromatography (TLC and open
column chromatography using silica gel added with 20% AgNO3 w/
w) and reverse-phase chromatography (C18, C5 and CN stationary
bonded phases HPLC columns with different solvent systems), but no
separation was achieved. The purity and homogeneity of all the
fractions from the chromatographic separation and the pure sapo-
genins were also checked by GC analyses.
were confirmed to be: 1: 3-O-
galactopyranosyl(1 / 2)- -D-glucuronopyranosyl soyasapogenol B
(soyasaponin I); 2: 3-O- -L-arabinopyranosyl(1 2)- -D-
glucopyranosyl(1 / 2)- -L-arabinopyranosyl hederagenin; 3: 3-O-
-L-arabinopyranosyl bayogenin; 4: 3-O- -D-glucuronopyranosyl-
28-O-[ -L-arabinopyranosyl(1 3)- -L-rhamnopyranoside]
medicagenic acid and 5: 3-O- -D-glucopyranosyl-28-O-[ -L-
arabinopyranosyl(1 / 3)- -L-rhamnopyranosyl(1 / 2)- -L-ara-
a-L-rhamnopyranosyl(1 / 2)-b-D-
b
a
/
b
a
a
b
a
/
a
b
a
a
a
binopyranoside] zanhic acid (Fig. 1).
4.3. Hydrolysis of saponins
5 mg of each pure saponin were separately treated with 30 mL of
2 N HCl in MeOH:H2O 1:1 under reflux for 8 and 30 h, respectively.
30 ml of H2O were than added, aglycones extracted with ethyl ac-
etate, the organic solution treated with anhydrous Na2SO4 and the
solvent eliminated under vacuum. The obtained aglycones were
dissolved in MeOH (2 ml) and all samples except sapogenins from 1,
were divided in two subsamples. One solution was treated with
CH2N2 and then silylated (with 0.2 ml of pyridin/hexamethyldisi-
lazane/chlorotrimethylsilane 2:1:1 and heated at 70 ꢀC for 10 min)
before GC injections. The other part of solution was directly sily-
lated and used to evaluate the amount of the methyl esters ob-
tained during the hydrolysis reaction. Sapogenins from 1 were
directly treated with the silylation mixture. Three independent
experiments were performed on each sample. Compound 1 was
also separately treated with 2 N HCl in CD3OD:H2O 1:1 and
EtOH:H2O 1:1 under reflux for 8 h and the obtained compounds
were treated as above and analysed by GC-MS.
4.5.1. 3
b
,22
b,24-trihydroxyolean-12(13)-en, soyasapogenol B (6)
White solid; C30H50O3, Mr 458; mp 252e253 ꢀC; [
a]
25 þ 98.8 (c
D
0.08 MeOH); IR (KBr) nmax 3420, 2950,1630, 1075 cmꢂ1; for 1H NMR
(300 MHz, CD3OD/CDCl3 2:1) and 13C NMR (75 MHz, CD3OD/CDCl3
2:1) see Table 1. Found: C, 78.7; H, 10.8. C30H50O3 requires: C, 78.5;
H, 11.0%.
4.5.1.1. 3b,22b,24-trihydroxyolean-12(13)-en trimethylsilyl (6a).
C
39H74O3Si3: GC-MS m/z (rel. int.) 674 (2) [M]þ; 584 (2) [Mꢂ90]þ;
481 (5) [M-90-CH2OSi(CH3)3]þ; 368 (7); 306 (92); 291 (86); 278
(18); 188 (40); 157 (70); 73 (100).
4.5.2. 3b,24-dihydroxyolean-12(13),21(22)-dien, soyasapogenol C
(11)
White solid; C30H48O2, Mr 440; mp 240e242 ꢀC; [
a
]
25 þ 33.8 (c
D
0.05 MeOH); IR (KBr) nmax 3435, 2948,1635,1075 cmꢂ1; for 1H NMR
Please cite this article in press as: Tava, A., et al., Artefact formation during acid hydrolysis of saponins from Medicago spp., Phytochemistry
(2017), http://dx.doi.org/10.1016/j.phytochem.2017.02.018

10
A. Tava et al. / Phytochemistry xxx (2017) 1e12
(300 MHz, CD3OD/CDCl3 2:1) and 13C NMR (75 MHz, CD3OD/CDCl3
2:1) see Table 1. Found: C, 81.7; H, 10.9. C30H48O2 requires: C, 81.8;
H, 11.0%.
yield) of crude sapogenin mixture. The sapogenin mixture was
submitted to a 400 ꢁ 55 mm, 40e60 m silica gel column (Merck).
m
Fractions were eluted with petroleum ether/CHCl3/acetic acid
(7:2:0.5) and checked by TLC developed with the some solvent
mixture, visualising the spots by spraying with MeOH/acetic an-
hydride/H2SO4 (10:1:1 v/v) followed by heating at 120 ꢀC. 12.1 mg
of pure compound 18 were obtained. The remaining fractions from
the silica gel column were further submitted to a preparative HPLC,
using a Perkin Elmer liquid chromatograph equipped with a LC250
binary pump and detected by UV monitoring at 215 nm. Fractions of
4.5.2.1. 3
b
,24-dihydroxyolean-12(13),21(22)-dien
trimethylsilyl
(11a). C36H64O2Si2: GC-MS m/z (rel. int.) 584 (2) [M]þ; 494 (4)
[Mꢂ90]þ; 391 (4) [M-90-CH2OSi(CH3)3]þ; 368 (6); 278 (11); 216
(95); 201 (32); 188 (28); 95 (48); 73 (100).
4.5.3. 4.5.3. 3
b
,22b
,24-trihydroxyolean-13(18)-en, soyasapogenol F
(13)
200
m
l (10 mg/ml of methanolic solution) were injected in a Dis-
25
White solid; C30H50O3, Mr 458; mp 312e314 ꢀC; [
a
]
-34.8 (c
covery C18 column (Supelco, 10 ꢁ 250 mm, 5
mm). Elution was
D
0.15 MeOH); IR (KBr) nmax 3448, 2951, 1645, 1080 cmꢂ1; for 1H NMR
(300 MHz, CD3OD/CDCl3 2:1) and 13C NMR (75 MHz, CD3OD/CDCl3
2:1) see Table 1. Found: C, 78.8; H, 10.8. C30H50O3 requires: C, 78.5;
H, 11.0%.
under isocratic condition of 35% MeOH, 65% H2O, 0.05% CF3COOH;
flow 2.0 ml/min. MeOH was removed under vacuum from the
collected fractions which were then freeze-dried. In total 8.5 mg of
pure zanhic acid 10 and 70.6 mg of compound 18 were obtained.
Compound 19 was obtained as methyl ester (7.2 mg) after prepar-
ative HPLC of the sapogenin mixture obtained after 30 h hydrolyses
of 100 mg of saponin 5. From a preparative HPLC a very small
amount (<1 mg) of compounds 21 and 22 were also obtained and
used to confirm their structure by GC/MS.
4.5.3.1. 3b,22b,24-trihydroxyolean-13(18)-en trimethylsilyl (13a).
C
39H74O3Si3: GC-MS m/z (rel. int.) 674 (2) [M]þ; 584 (2) [Mꢂ90]þ;
494 (2) [M-90-90]þ; 481 (7) [M-90-CH2OSi(CH3)3]þ; 391 (9) [M-90-
CH2OSi(CH3)3-90]þ; 306 (5); 278 (26); 203 (65); 157 (100); 73 (84).
4.5.4. 3
b,24-dihydroxy-22b-methoxyolean-13(18)-en,
4.6.1. 2
acid (10)
White solid; C30H46O7, Mr 518; mp > 320 ꢀC; [
b,3b,16a-trihydroxyolean-12(13)-en-23,28 dioic acid, zanhic
soyasapogenol D (15)
25
White solid; C31H52O3, Mr 472; mp 289e291 ꢀC; [
a
]
-51.8 (c
D
a
]
D
25 þ65.8 (c 0.52
0.29 MeOH); IR (KBr) nmax 3445, 2940,1630,1075 cmꢂ1; for 1H NMR
(300 MHz, CD3OD/CDCl3 2:1) and 13C NMR (75 MHz, CD3OD/CDCl3
2:1) see Table 1. Found: C, 78.6; H, 10.8. C31H52O3 requires: C, 78.8;
H, 11.1%.
MeOH); IR (KBr) nmax 3450, 2945, 1715, 1630, 1085 cmꢂ1; for 1H
NMR (300 MHz, CD3OD) and 13C NMR (75 MHz, CD3OD) see Table 2.
ESI-MS-MS m/z (rel. int.) 1035.2 (9) [2M-H]ꢂ; 517.1 (100) [M-H]ꢂ;
MS2: 499.0 (100) [M-H-H2O]ꢂ; 471.0 (12) [M-H-H2O-CO]ꢂ; 455.0
(16) [M-H-H2O-CO2]ꢂ; 453.1 (13) [M-H-2H2O-CO]ꢂ; 437.0 (25) [M-
H-2H2O-CO2]ꢂ. Found: C, 69.8; H, 8.7%. C30H46O7 requires: C, 69.5;
H, 8.9%.
4.5.4.1. 3b,24-dihydroxy-22b -methoxyolean-13(18)-en trimethylsilyl
(15a). C37H68O3Si2: GC-MS m/z (rel. int.) 616 (3) [M]þ; 526 (2) [M-
90]þ; 426 (9) [M-90-CH2OSi(CH3)3]þ; 391 (8) [M-90-CH2OSi(CH3)3-
CH3OH]þ; 278 (29); 203 (47); 187 (46); 99 (92); 73 (100).
Dimethyl ester. White solid; C32H50O7, Mr 546; mp 261e263 ꢀC;
25
[a
]
þ26.0 (c 0.62 MeOH). Found: C, 70.1; H, 9.4%. C32H50O7 re-
D
quires: C, 70.3; H, 9.2%.
4.5.4.2. 3b,24-dihydroxy-22b-deuteromethoxyolean-13(18)-en tri-
methylsilyl (15b). C37H65D3O3Si2: GC-MS m/z (rel. int.) 619 (2) [M]þ;
530 (2) [M-89]þ; 426 (7) [M-90-CH2OSi(CH3)3]þ; 391 (6) [M-90-
CH2OSi(CH3)3-CD3OH]þ; 278 (25); 203 (28); 187 (35); 147 (45);
102 (87); 73 (100).
4.6.1.1. 2
b,3
b,16a-trihydroxyolean-12(13)-en-23,28
dioic
acid
dimethyl ester trimethylsilyl (10a). C41H74O7Si3: GC-MS m/z (rel.
int.) 762 (3) [M]þ; 703 (2) [M-59]þ; 672 (10) [M-90]þ; 613 (5) [M-
90-59]þ; 582 (3) [M-90-90]þ; 523 (3) [M-90-90-59]þ; 411 (2); 350
(10); 260 (93); 201 (100); 187 (20); 173 (38); 147 (54); 133 (43); 73
(58).
4.5.4.3. 3b,24-dihydroxy-22b-ethoxyolean-13(18)-en trimethylsilyl
(15c). C38H70O3Si2: GC-MS m/z (rel. int.) 630 (2) [M]þ; 540 (2) [M-
90]þ; 437 (7) [M-90-CH2OSi(CH3)3]þ; 391 (7) [M-90-CH2OSi(CH3)3-
CH3CH2OH]þ; 278 (23); 203 (42); 187 (31); 147 (40); 113 (92); 73
(100).
4.6.2. 2
b,3b,16a-trihydroxyolean-13(18)-en-23,28 dioic acid (18)
White solid; C30H46O7, Mr 518; mp > 320 ꢀC; [
a]
25 þ32.0 (c 0.51
D
MeOH); IR (KBr) nmax 3455, 2943, 1720, 1634, 1090 cmꢂ1; for 1H
NMR (300 MHz, CD3OD) and 13C NMR (75 MHz, CD3OD) see Table 2.
ESI-MS-MS m/z (rel. int.) 1035.3 (13) [2M-H]ꢂ; 517.1 (100) [M-H]ꢂ;
MS2: 499.0 (86) [M-H-H2O]ꢂ; 473.0 (81) [M-H-CO2]ꢂ; 471.2 (13)
[M-H-H2O-CO]ꢂ; 455.0 (100) [M-H-H2O-CO2]ꢂ; 453.1 (19) [M-H-
2H2O-CO]ꢂ; 437.0 (11) [M-H-2H2O-CO2]ꢂ. Found: C, 69.2; H, 8.7.
4.5.5. 3b,22b,24-trihydroxyolean-18(19)-en, soyasapogenol H (16)
For 1H NMR (300 MHz, CD3OD/CDCl3 2:1) and 13C NMR (75 MHz,
CD3OD/CDCl3 2:1) see Table 1.
4.5.5.1. 3b,22b,24-trihydroxyolean-18(19)-en trimethylsilyl (16a).
C
39H74O3Si3: GC-MS m/z (rel. int.) 674 (7) [M]þ; 585 (15) [M-89]þ;
481 (4) [M-89-CH2OSi(CH3)3]þ; 391 (6) [M-89-CH2OSi(CH3)3-90]þ;
292 (9); 277 (9); 202 (18); 187 (29); 175 (48); 147 (40); 73 (100).
C
30H46O7 requires: C, 69.5; H, 8.9%.
Dimethyl ester. White solid; C32H50O7, Mr 546; mp 265e268 ꢀC;
[a]
25 -5.4 (c 0.65 MeOH). Found: C, 70.1; H, 9.3. C32H50O7 requires:
D
4.6. Hydrolysis of saponin 5 and purification of compounds 10, 18
and 19
C, 70.3; H, 9.2%.
Saponin 5 (350 mg, 321.1
mmol) was treated with 200 ml of 2 N
4.6.2.1. 2
b
,3
b,16
a-trihydroxyolean-13(18)-en-23,28
dioic
acid
HCl in MeOH:H2O 1:1 under reflux for 8 h. MeOH was removed
under reduced pressure and aglycones were extracted by using
ethyl acetate (3 ꢁ 100 ml). The organic solution was treated with
anhydrous Na2SO4 and the solvent removed under reduced pres-
dimethyl ester trimethylsilyl (18a). C41H74O7Si3: GC-MS m/z (rel.
int.) 762 (14) [M]þ; 703 (2) [M-59]þ; 672 (9) [M-90]þ; 613 (12) [M-
90-59]þ; 582 (5) [M-90-90]þ; 523 (11) [M-90-90-59]þ; 463 (5) [M-
90-90-59-60]þ; 411 (7); 335 (20); 321 (58); 261 (48); 247 (41); 201
(39); 187 (68); 173 (67); 147 (100); 133 (79); 73 (79).
sure to yield 152.2 mg (293.9 mmol as zanhic acid equivalent, 91.5%
Please cite this article in press as: Tava, A., et al., Artefact formation during acid hydrolysis of saponins from Medicago spp., Phytochemistry
(2017), http://dx.doi.org/10.1016/j.phytochem.2017.02.018

A. Tava et al. / Phytochemistry xxx (2017) 1e12
11
4.6.3. 2
ester (19)
White solid; C31H48O7, Mr 532; mp > 320 ꢀC; [
b
,3
b
,16
a
-trihydroxyolean-28,13
b
-olide-23-oic acid methyl
4.7.6. 2
b,3b-dihydroxyolean-23-oic acid-28,13b-olide methyl ester
trimethylsilyl (29a)
a
]
25 þ15.2 (c 0.89
C
37H64O6Si2: GC-MS m/z (rel. int.) 660 (2) [M]þ; 570 (6) [M-90]þ;
D
MeOH); IR (KBr) nmax 3450, 2946, 1770, 1105 cmꢂ1; for 1H NMR
(300 MHz, CD3OD) and for 13C NMR (75 MHz, CD3OD) see Table 2.
ESI-MS-MS m/z (rel. int.) 1063.4 (100) [2M-H]ꢂ; 531.0 (17) [M-H]ꢂ;
MS2: 499.0 (100) [M-H2O-CH3]ꢂ; 471.0 (3) [M-H2O-CH3-CO]ꢂ;
455.0 (2) [M-H2O-CH3-CO2]ꢂ; 453.1 (3) [M-2H2O-CH3-CO]ꢂ; 441.0
(5) [M-H2O-CO-CO2]ꢂ. Found: C, 70.3; H, 9.2. C31H48O7 requires: C,
69.9; H, 9.1%.
511 (18) [M-90-59]þ; 429 (10); 411 (6); 321 (19); 275 (12); 218 (27);
189 (25); 173 (28); 147 (55); 73 (100).
Acknowledgements
This research was supported by grant from ‘Consiglio per la
Ricerca in Agricoltura e l’Analisi dell’Economia Agraria’ (CREA),
ꢀ
Italy, and by the Italian ‘Ministero dell’Istruzione, dell’Universita e
4.6.3.1. 2b,3b,16a-trihydroxyolean-28,13b-olide-23-oic acid methyl
della Ricerca’.
ester trimethylsilyl (19a). C40H72O7Si3: GC-MS m/z (rel. int.) 748 (4)
[M]þ; 689 (2) [M-59]þ; 658 (7) [M-90]þ; 599 (15) [M-90-59]þ; 509
(6) [M-90-59-90]þ; 411 (10); 334 (22); 321 (35); 306 (22); 275 (32);
244 (23); 231 (30); 187 (48); 173 (59); 147 (100); 133 (60); 75 (96).
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.phytochem.2017.02.018.
4.6.4. 2b,3b-dihydroxyolean-12(13),15(16)-dien-23,28-dioic acid
(21)
References
C
30H44O6, Mr 500.
Aneja, R., Davies, A.P., Knaggs, J.A., 1975. Formation of a 3,5-cyclocholestan-6a-yl
derivative in a nucleophilic substitution reaction of cholesterol. Tetrahedron
Lett. 12, 1033e1036.
4.6.4.1. 2 ,3 -dihydroxyolean-12(13),15(16)-dien-23,28-dioic
b
b
acid
dimethyl ester trimethylsilyl (21a). C38H64O6Si2: GC-MS m/z (rel.
int.) 762 (4) [M]þ; 613 (2) [M-59]þ; 582 (2) [M-90]þ; 523 (7) [M-90-
59]þ; 463 (2) [M-90-59-60]þ; 433 (2) [M-90-59-90]þ; 411 (8); 321
(22); 260 (53); 201 (100); 187 (28); 173 (32); 147 (38); 133 (57).
Avato, P., Bucci, R., Tava, A., Vitali, C., Rosato, A., Bialy, Z., Jurzysta, M., 2006. Anti-
microbial activity of saponins from Medicago sp.: structure-activity relation-
ship. Phytother. Res. 20, 454e457.
Baxter, R., Price, K.R., Fenwick, G.R., 1990. Sapogenin structure: analysis of the 13C-
and 1H-NMR spectra of soyasapogenol B. J. Nat. Prod. 53, 289e302.
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,
3185e3192.
4.6.5. 2
29H44O4, Mr 456. UV (MeOH) lmax (log ε) 235 (2.85), 243 (2.84),
253 (2.66).
b,3b-dihydroxyolean-12(13),17(18)-dien-23-oic acid (22)
C
Biazzi, E., Carelli, M., Tava, A., Abbruscato, P., Losini, I., Avato, P., Scotti, C.,
Calderini, O., 2015. CYP72A67 catalyzes a key oxidative step in Medicago trun-
catula hemolytic saponin biosynthesis. Mol. Plant 8, 1493e1506.
Brawn, P.R., Lindner, N.M., Miller, J.M., Telling, G.M., 1981. A gas chromatographic
method for the determination of medicagenic acid in lucerne (alfalfa) leaf
protein concentrate. J. Sci. Food Agric. 32, 1157e1162.
Budzikiewicz, H., Wilson, J.M., Djerassi, C., 1963. Mass spectrometry in structural
and stereochemical problems. XXXII. Pentacyclic triterpenes. J. Am. Chem. Soc.
85, 3688e3699.
4.6.5.1. 2b,3b-dihydroxyolean-12(13),17(18)-dien-23-oic acid methyl
ester trimethylsilyl (22a). C36H62O4Si2: GC-MS m/z (rel. int.) 614
(23) [M]þ; 524 (4) [M-90]þ; 465 (5) [M-90-59]þ; 321 (5); 202 (55);
190 (48); 187 (28); 173 (22); 147 (59); 133 (38).
Cadenas, R.A., Gelpi, M.E., Mosettig, J., 2005. Nucleosteroids: synthesis of purinyl-
steroid derivatives under Mitsunobu reaction conditions. J. Hetrocyclic Chem.
42, 1e3.
Carelli, M., Biazzi, E., Panara, F., Tava, A., Scaramelli, L., Porceddu, A., Graham, N.,
Odoardi, M., Piano, E., Arcioni, S., May, S., Scotti, C., Calderini, O., 2011. Medicago
truncatula CYP716A12 is a multifunctional oxidase involved in the biosynthesis
of hemolytic saponins. Plant Cell 23, 3070e3081.
Ding, K., Sun, Y.S., Tian, W.S., 2011. Total synthesis of ( )-spiniferin-1 via a poly-
fluoroalkanosulfonyl fluoride induced homoallilic carbocation rearrangement
reaction. J. Org. Chem. 76, 1495e1498.
Fukushima, E.O., Seki, H., Sawai, S., Suzuki, M., Ohyama, K., Saito, K., Muranaka, T.,
2013. Combinatorial biosynthesis of legume natural and rare triterpenoids in
engineered yeast. Plant Cell Physiol. 54, 740e749.
4.7. GC/MS data of artefact compounds from acid hydrolysis of
saponins 2, 3 and 4
4.7.1. 2b,23-dihydroxyolean-13(18)-en-23-oic acid methyl ester
trimethylsilyl (24a)
C
37H66O4Si2: GC-MS m/z (rel. int.) 630 (2) [M]þ; 540 (3) [M-
90]þ; 481 (1) [M-90-59]þ; 278 (15); 248 (10); 201 (21); 189 (51);
173 (21); 148 (100); 133 (37).
Heftmann, E., Lundin, R.E., Haddon, W.F., Peri, I., Mor, U., Bondi, A., 1979. High-
pressure liquid chromatography, nuclear magnetic resonance and mass spectra
of biosynthetic soyasapogenols. J. Nat. Prod. 42, 410e416.
Huhman, D.V., Berhow, M.A., Sumner, L.W., 2005. Quantification of saponins in
aerial and subterranean tissues of Medicago truncatula. J. Agric. Food Chem. 53,
1914e1920.
4.7.2. 2b,23-dihydroxyolean-28,13b-olide trimethylsilyl (25a)
C
36H64O4Si2: GC-MS m/z (rel. int.) 616 (2) [M]þ; 526 (10) [M-
90]þ; 436 (6) [M-90-90]þ; 235 (8); 200 (20); 187 (23); 173 (12); 147
(58); 135 (18); 73 (100).
Ireland, P.A., Dziedzic, S.Z., 1986. Effect of hydrolysis on sapogenin release in soya.
J. Agric. Food Chem. 34, 1037e1041.
4.7.3. 2
trimethylsilyl (26a)
b,3b,23-trihydroxyolean-13(18)-en-23-oic acid methyl ester
Jurzysta, M., Jurzysta, A., 1978. Gas-liquid chromatography of trimethylsilyl ethers of
soya sapogenols and medicagenic acid. J. Chromatogr. 148, 517e520.
Jurzysta, M., 1984. Transformation of soyasapogenol B into soyasapogenols C, D and
F under acidic conditions. In: Proc. 14th Int. Symp. Nat. Prod, p. 127.
Kapusta, I., Janda, B., Stochmal, A., Oleszek, W., 2005. Determination of saponins in
aerial parts of barrel medic (Medicago truncatula) by liquid chromatography-
electrospray ionization/mass spectrometry. J. Agric. Food Chem. 53, 7654e
7660.
C
40H74O5Si3: GC-MS m/z (rel. int.) 718 (2) [M]þ; 659 (1) [M-60]þ;
628 (5) [M-90]þ; 538 (6) [M-90-90]þ; 525 (12); 465 (7); 275 (12);
261 (10); 191 (40); 173 (20); 147 (47); 133 (37); 73 (100).
4.7.4. 2
39H72O5Si3: GC-MS m/z (rel. int.) 704 (1) [M]þ; 614 (9) [M-90]þ;
524 (6) [M-90-90]þ; 511 (23); 191 (28); 147 (57); 73 (100).
b,3b,23-trihydroxyolean-28,13b-olide trimethylsilyl (27a)
C
Kinjo, J., Miyamoto, I., Miurakami, K., Kida, K., Tomimatsu, T., Yamasaky, M.,
Nohara, T., 1985. Oleanane sapogenols from Puerare radix. Chem. Pharm. Bull.
33, 1293e1296.
Kojima, H., Ogura, H., 1989. Configurational studies on hydroxyl groups at C-2, 3 and
23 or 24 of oleanane and ursane-type triterpenes by NMR spectroscopy.
Phytochemistry 28, 1703e1710.
Kubota, T., Kitatani, H., Hinoh, H., 1969. Isomerisation of quillaic acid and echino-
cystic acid with hydrochloric acid. Tetrahedron Lett. 10, 771e774.
Mahato, S.B., 1991. Triterpenoid saponins from Medicago hyspida. Phytochemistry
30, 3389e3393.
4.7.5. 2
ester trimethylsilyl (28a)
b,3b-dihydroxyolean-13(18)-en-23,28-dioic acid methyl
C
38H66O6Si3: GC-MS m/z (rel. int.) 674 (2) [M]þ; 584 (5) [M-90]þ;
525 (12) [M-90-59]þ; 465 (7) [M-90-59-60]þ; 411 (3); 335 (5); 321
(19); 261 (11); 189 (45); 147 (38); 133 (34); 73 (100).
Please cite this article in press as: Tava, A., et al., Artefact formation during acid hydrolysis of saponins from Medicago spp., Phytochemistry
(2017), http://dx.doi.org/10.1016/j.phytochem.2017.02.018

12
A. Tava et al. / Phytochemistry xxx (2017) 1e12
Mahato, S.B., Kundu, A.P., 1994. 13C NMR spectra of pentacyclic triterpenoids e a
compilation and some salient features. Phytochemistry 37, 1517e1575.
Martinez, A., Perojil, A., Rivas, F., Parra, A., Garcia-Granados, A., Fernandez-Vivas, A.,
2015. Biotransformation of oleanolic and maslinic methyl esters by Rhizomucor
miehei CECT 2749. Phytochemistry 117, 500e508.
Marx Young, M.C., Potomati, A., Chu, E.P., Haraguchi, M., Yamamoto, M., Kawano, T.,
1997. 13C NMR analysis of monodesmosidic saponins from Gomphrena macro-
cephala. Phytochemistry 46, 1267e1270.
and interconversion. J. Sci. Food Agric. 37, 1027e1034.
Rao, D., Bories, G., 1987. Simple gas chromatographic method for the determination
of medicagenic acid in alfalfa (Medicago sativa). J. Chromatogr. 410, 169e175.
Tava, A., Oleszek, W., Jurzysta, M., Berardo, N., Odoardi, M., 1993. Alfalfa saponins
and sapogenins: isolation and quantification in two different cultivars. Phy-
tochem. Anal. 4, 269e274.
Tava, A., Pecetti, L., 1998. Hemolytic activity and saponin content in lucerne (Med-
icago sativa complex) genotypes. J. Gen. Breed. 52, 33e37.
Massiot, G., Lavaud, C., Guillaume, D., Le Men-Olivier, L., 1988. Reinvestigation of
sapogenins and prosapogenins from alfalfa (Medicago sativa). J. Agric. Food
Chem. 36, 902e909.
Tava, A., Odoardi, M., Oleszek, W., 1999. Seasonal changes of saponin content in five
alfalfa (Medicago sativa) cultivars. Agric. Medit. 129, 111e116.
Tava, A., Chiari, M., Oleszek, W., 2000. Separation of alfalfa (M. sativa L.) saponins as
their borate complexes by capillary electrophoresis (Chapter 5). In: Oleszek, W.,
Marston, A. (Eds.), Saponins in Food, Feedstuffs and Medicinal Plants. Kluwer
Academic Publishers, Dordrecht, The Netherland, pp. 43e56.
Tava, A., Mella, M., Bialy, Z., Jurzysta, M., 2003. Stability of saponins in alcoholic
solutions: ester formation as artifacts. J. Agric. Food Chem. 51, 1797e1800.
Tava, A., Mella, M., Avato, P., Argentieri, M.P., Bialy, Z., Jurzysta, M., 2005. Triterpe-
noid glycosides from the leaves of Medicago arborea L. J. Agric. Food Chem. 53,
9954e9965.
Tava, A., Mella, M., Avato, P., Biazzi, E., Pecetti, L., Bialy, Z., Jurzysta, M., 2009. New
triterpenic saponins from the aerial parts of Medicago arabica (L.) Huds. J. Agric.
Food Chem. 57, 2826e2835.
Massiot, G., Dijoux, M.G., Lavaud, C., 1996. Saponins and artifacts. In: Waller, G.R.,
Yamasaki, K. (Eds.), Saponins Used in Food and Agricolture. Plenum Press, New
York, pp. 183e192.
~
Moses, T., Pollier, J., Almagro, L., Buyst, D., Van Montagu, M., Pedreno, M.A.,
Martins, J.C., Thevelein, J.M., Goossens, A., 2014. Combinatorial biosynthesis of
sapogenins and saponins in Saccharomices cerevisiae using a C-16a hydroxylase
from Buplerum falcatum. Proc. Natl. Acad. Sci. U. S. A. 111, 1634e1639.
Nes, W.D., Benson, M., Heftmann, E., 1981. The location of the methylol group in
sapogenol C and erythrodiol and its biosynthetic significance. Phytochemistry
20, 2299e2300.
Nowacka, J., Oleszek, W., 1994. Determination of alfalfa (Medicago sativa) saponins
by high-performance liquid chromatography. J. Agric. Food Chem. 42, 727e730.
Pecetti, L., Tava, A., Romani, M., De Benedetto, M.G., Corsi, P., 2006. Variety and
environment effects on the dynamics of saponins in lucerne (Medicago sativa
L.). Eur. J. Agron. 25, 187e192.
Tava, A., Scotti, C., Avato, P., 2011. Biosynthesis of saponins in the genus Medicago.
Phytochem. Rev. 10, 459e469.
Witkowska, H.E., Bialy, Z., Jurzysta, M., Waller, G.R., 2008. Analysis of saponin
mixtures from alfalfa (Medicago sativa L.) roots using mass spectrometry with
MALDI techniques. Nat. Prod. Comm. 3, 1395e1410.
Price, K.R., Fenwick, G.R., Jurzysta, M., 1986. Soyasapogenols e separation, analysis
Please cite this article in press as: Tava, A., et al., Artefact formation during acid hydrolysis of saponins from Medicago spp., Phytochemistry
(2017), http://dx.doi.org/10.1016/j.phytochem.2017.02.018