Regioselective synthesis of 7-O-esters of the flavonolignan silibinin and SARs lead to compounds with overadditive neuroprotective effects

Article abstract of DOI:10.1016/j.ejmech.2018.01.036

A series of neuroprotective hybrid compounds was synthesized by conjugation of the flavonolignan silibinin with natural phenolic acids, such as ferulic, cinnamic and syringic acid. Selective 7-O-esterfication without protection groups was achieved by appl

Full text of DOI:10.1016/j.ejmech.2018.01.036

Accepted Manuscript
Regioselective synthesis of 7-O-esters of the flavonolignan silibinin and SARs lead to
compounds with overadditive neuroprotective effects
Simon Schramm, Guozheng Huang, Sandra Gunesch, Florian Lang, Judit Roa, Petra
Högger, Raimon Sabate, Pamela Maher, Michael Decker
PII:
S0223-5234(18)30049-7
10.1016/j.ejmech.2018.01.036
EJMECH 10113
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 11 December 2017
Revised Date: 10 January 2018
Accepted Date: 11 January 2018
Please cite this article as: S. Schramm, G. Huang, S. Gunesch, F. Lang, J. Roa, P. Högger, R. Sabate,
P. Maher, M. Decker, Regioselective synthesis of 7-O-esters of the flavonolignan silibinin and SARs
lead to compounds with overadditive neuroprotective effects, European Journal of Medicinal Chemistry
(2018), doi: 10.1016/j.ejmech.2018.01.036.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: www.elsevier.com/locate/ejmech
Regioselective synthesis of 7-O-esters of the flavonolignan silibinin and
SARs lead to compounds with overadditive neuroprotective effects
Simon Schramm1, Guozheng Huang1, Sandra Gunesch1, Florian Lang2, Judit Roa3, Petra
Högger2, Raimon Sabate3, Pamela Maher4 and Michael Decker1
1Pharmazeutische und Medizinische Chemie, Institut für Pharmazie und Lebensmittelchemie, Julius-Maximilians-Universität Würzburg, Am
Hubland, D-97074 Würzburg, Germany
2Klinische Pharmazie, Institut für Pharmazie und Lebensmittelchemie, Julius-Maximilians-Universität Würzburg, Am Hubland, D-97074
Würzburg, Germany
3Department of Pharmacy and Pharmaceutical Technology and Physical-Chemistry, School of Pharmacy, and Institute of Nanoscience and
Nanotechnology (IN2UB), University of Barcelona, E-08028, Barcelona, Spain
4The Salk Institute for Biological Studies, 10010 N. Torrey Pines Road, La Jolla, California 92037, USA
ABSTRACT
A series of neuroprotective hybrid compounds was synthesized by conjugation of the
flavonolignan silibinin with natural phenolic acids, such as ferulic, cinnamic and syringic
acid. Selective 7-O-esterfication without the need for protection groups was achieved by
applying the respective acyl chlorides. Sixteen compounds were obtained and SARs were
established by evaluating antioxidative properties in the physicochemical FRAP assay, as
well as in a cell-based neuroprotection assay using murine hippocampal HT-22 cells.
Despite weak activities in the FRAP assay, esters of the α,β-unsaturated acids showed
pronounced overadditive effects at low concentrations greatly exceeding the effects of
equimolar mixtures of silibinin and the respective acids in the neuroprotection assay.
Cinnamic and ferulic acid esters (5a and 6a) also showed overadditive effects regarding
inhibition of microglial activation, PC12 cell differentiation, in vitro ischemia as well as
anti-aggregating abilities against Aβ42 peptide and τ protein. Remarkably, the esters of
ferulic acid with silybin A and silybin B (11a and 11b) showed a moderate but significant
difference in both neuroprotection and in their anti-aggregating capacities. The results
demonstrate that non-toxic natural antioxidants can be regioselectively connected as
esters with medium-term stability exhibiting very pronounced overadditive effects in a
portfolio of biological assays.
Article history:
Received
Revised
Accepted
Available online
Keywords:
Silybin
Esterification
HT-22 cells
FRAP assay
Stability
1. Introduction
One of the key processes in the pathophysiology of
Alzheimer’s disease (AD) is neurotoxic oxidative
stress [1]. The formation of reactive oxygen species
contributes to and enhances neuronal cell death.
Therefore antioxidant agents might help to prevent or
slow disease progression. For that reason antioxidant
compounds are often incorporated into hybrid
molecules [2].
There are several reports that standardized extracts of
milk thistle (Silybum marianum (L.) GAERTN.),
Figure 1. Structure of silibinin (1a, 1b), silibinin is a
diastereomeric mixture of silybin A 1a and silybin B 1b

ACCEPTED MANUSCRIPT
containing the silymarin complex as its bioactive
principle therapeutically applied for hepatoprotection,
as well as its main component silibinin (1a and 1b,
Figure 1) and chemical derivatives of silibinin exhibit
neuroprotective properties. These findings are
ambiguous though, depending on which exact
chemical structure was used and which biological
assay applied [3-6]. The flavonolignan silibinin (1) is
a natural antioxidant that contains five hydroxy
groups, three phenolic, a primary and a secondary
one. In several publications [7-9] the influence of the
individual hydroxy groups of silibinin (1) on its
antioxidant and radical scavenging properties were
studied in detail (Figure 2A). The findings led to the
conclusion that the phenolic 20-OH-moiety is
essential for the compounds radical-scavenging
properties, whereby the 7-OH-group may possess a
pro-oxidant character [8]. Therefore blocking the 7-
OH-group by introduction of another radical-
scavenging moiety may improve the antioxidant
properties of silibinin (1).
Hence
a
compound unifying the antioxidant
properties of both silibinin (1) and ferulic acid (2)
may potentially exhibit superior neuroprotective and
antioxidant effects, similar to the very recently
published hybrid antioxidants from Vavříková et al.
[15], in which for example ascorbic acid and trolox
alcohol were combined at the 23-position hydroxyl
group of silibinin over dodecanedioic acid spacers as
diesters.
Formation of esters can be used to co-apply both
antioxidant entities with improved bioavailability due
to higher lipophilicity compared to silibinin and
protection of the 7-position. Therefore, we decided to
synthesize a library of 7-O-esters of silibinin (1) with
derivatives of ferulic acid (2), using substituted and
non-substituted cinnamic, dihydro cinnamic and
benzoic acids (for overview of target structures - see
Figure 2B)
to
establish
structure-activity-
relationships. Thus, the study is designed to reveal the
influence of different numbers of hydroxy- and
methoxy-groups as aromatic substituents as well as
the Michael system on neuroprotective and
antioxidant effects. In all cases, we used natural
products or simple derivatives thereof.
Ferulic acid (2, Figure 2A) represents a naturally
occurring phenolic acid [10], that has been shown to
have a number of remarkably positive effects on AD
pathology such as disaggregating effects on amyloid
structures and memory improving effects in mice [11,
12]. Furthermore, ferulic acid (2) is a potent radical
scavenger, as measured in the physicochemical
oxygen-radical absorbance capacity (ORAC) assay
[13], but it lacks neuroprotective activity in vitro after
impairment of murine hippocampal HT-22 cells with
tert-butyl hydroperoxide [14].
Regioselective esterification of silibinin (1) is a
challenging task [16]. A number of silibinin esters
have been described before (for an overview of
prepared silibinin derivatives see [17]). Selective
esterification in the 23-position under Mitsunobu
conditions was reported previously by Wang et al. [5]
and by our group in preparation of a multipotent
silibinin-tacrine co-drug [18]. Also, Gažák et al.
Figure 2. A) Postulated contribution of individual OH-groups of silibinin (1) on antioxidant activity [7-9] and structure of ferulic acid
(2). B) Structures of target 7-O-esters of silibinin (1).

ACCEPTED MANUSCRIPT
prepared several silibinin galloyl esters in positions 3,
assay and d) for their inhibition of Aβ peptide and τ
7, 20 and 23 using different protection/deprotection
strategies [19] and acyl chlorides for the preparation
of 7-O- and 23-O-esters [20].
protein aggregation [28, 29].
Concerning the stability of the target ester
compounds, the ester bond may be hydrolyzed in
solution or by esterases [30] present in various
tissues. But there are reports [17, 31] suggesting
inhibition of esterases by silibinin (1) similar to other
flavonoids [32, 33]. In fact, our work as well as that
from other groups [18b, 19] have shown remarkable
stability of various silibinin ester groups, probably
due to steric hindrance with regard to enzyme
accessibility compared to the 23-position, which is
more prone to hydrolytic attack. The latter has been
utilized for the development of codrugs [18].
Nevertheless, determination of the stability of the
compounds is important to distinguish their effects
from the effects of the individual components or a
mixture thereof.
Furthermore, it had been reported that the pure
diastereomers of silibinin (1), i. e. silybin A (1a) and
silybin B (1b), their 2,3-dehydro-derivatives and
other synthetic derivatives show both distinct
metabolic profiles [21] and biological effects, such as,
influence on amyloid-β (Aβ) formation [22],
antiangiogenic effects [19], antiproliferative effects in
human carcinoma cells [23], modulation of endocrine
functions [24] and anti-aging effects [25]. Therefore,
potentially differing effects of pure silybin A and
silybin B esters were also investigated.
In order to evaluate neuroprotective activity of the
compounds synthesized, a cell-based model system
that uses HT-22 hippocampal neurons was applied
[14, 18, 26]. Administration of antioxidants such as
flavonoids can effectively prevent oxidative neuronal
cell death in this cell line [14, 26]. Additionally,
putative innate neurotoxic effects of the compounds
were studied on HT-22 cells in cytotoxicity studies.
The interplay of a) improved compound properties
compared to the poor pharmacokinetics of silibinin,
b) the co-administration of two pharmacologically
active compounds, c) the overadditive antioxidant
effects of esters (cf. below) and d) their medium-term
hydrolysis to known and non-toxic antioxidants with
described pharmacokinetics, might make ester-
Further biological evaluation of the most promising
compounds was accomplished by screening activities
Table 1. Reaction conditions and the structures of synthesized ester derivatives (4a-e, 6a) with regard to regioselectivity
Entry
Reaction conditions
Product
No.
6a
R1
H
R2
H
R3
H
R4
H
1
Ferulic acid (2), PPh3, DIAD, THF, 10 oC - r.t., 20 hr; or (i).
Ferulic acid, oxalyl chloride, CH2Cl2, r.t. 30 min, (ii). pyridine,
r.t. 1 hr.
2
3,4-O-(1-methylethylidene)caffeic acid (3), PPh3, DIAD, THF,
H
H
4a
10 o C - r.t., 40 hr.
4
5
6
Acetyl chloride, triethylamine, CH2Cl2/THF, r.t. 1 hr.
Acetic anhydride, BF3·Et2O, THF, 0 – r.t. 18 hr; [31, 34]
(i). 4, 4’-Dimethoxytrityl chloride, pyridine, 50 o C. 20 hr; (ii).
acetic anhydride, pyridine, r.t. 18 hr; (iii). 5% formic acid in
CH2Cl2, r.t. 30 min. [4]
H
H
H
4b
4c
4d
Ac
H
Ac
Ac
H
H
Ac
H
Ac
7
(i). (4-O-acetyl)ferulic acid, oxalyl chloride, CH2Cl2, r.t. 30
H
H
H
4e
min; (ii). pyridine, r.t. 1 hr.
in five other assays that focus on various aspects of
AD pathophysiology [27]. Compounds were tested a)
in an in vitro ischemia model, b) for inhibition of
microglial activation, c) in a PC12 cell differentiation
codrugs a novel approach in the development of
silibinin derivatives.
2. Results and discussion

ACCEPTED MANUSCRIPT
acetyl)feruloylsilibinin (4e, Table 1) according to the
2.1 Chemistry
1D and 2D NMR data.
For the synthesis of the 7-O-esters we investigated
and probed several esterfication methods to achieve
the desired regioselectivity in a consistent manner
without the use of time consuming protecting group
strategies. One of the first methods we tested was
Mitsunobu esterfication.
Because of the inconsistent regioselectivity as well as
the formation of dehydration products (i.e.
hydnocarpin D and derivatives – we have previously
optimized dehydration conditions to get semi-
synthetic access to such derivatives under Mitsunobu
conditions [36]) of the Mitsunobu reaction, we started
testing 7-esterification using respective acyl chlorides
under basic conditions.
To our surprise, applying similar Mitsunobu
conditions as described in references [5] and [18] to
ferulic acid (2) and silibinin (1) led to 7-O-
feruloylsilibinin (6a, Table 1) rather than the expected
23-O-feruloylsilibinin as confirmed by 1D and 2D
NMR. On the contrary, using an acetonide derivative
of caffeic acid 3, the only isolated product was the
23-esterified product, i. e. 23-O-[3,4-O-(1-
methylethylidene)caffeoyl]silibinin (4a, Table 1)
which was consistent with the previous reports [5,
18], together with unconsumed acid 3. The presence
or absence of downfield shifts of H-23 signals clearly
indicated esterification on this position. The NMR
data of compounds 6a and 4a were also compared
with that of 7-O-acetylsilibinin (4b), 23-O-
acetylsilibinin (4c) [31, 34] and 3,5,7,20-
tetraacetylsilibinin (4d) [4]. The signals of free
respective OH groups of silibinin derivatives in
CDCl3 were determined by 1D and 2D NMR, such as
the signal of 3-OH appears at ~3. 5 ppm, while the
signal of 23-OH appears at ~ 2.0 ppm (Table 2). The
presence or absence of respective OH group’s signals
also helped to determine the position of esterification.
Similar to conditions applied before [19, 20],
compound 6a could also be obtained using feruloyl
chloride, which was prepared in situ without
protection of the hydroxyl group of ferulic acid
(Table 3). As this alternative method produces less
byproducts and does not necessitate tedious
chromatographic purification as required for
a
Mitsunobu reaction, we decided to use acyl chlorides
for the syntheses of the 7-O-esters. The optimized
procedure of in situ preparation of the acyl chloride,
using oxalyl chloride and DMF in dichloromethane,
and adding the resulting solution to a solution of
silibinin (1) and 4.5 equivalents triethylamine in
tetrahydrofuran, gave the 7-O-esters 5-8. In most
cases simple column chromatography after workup
was sufficient to obtain the desired compounds in
good purity (overview of synthesized compounds in
Table 3).
To investigate a putative differing effect of the
individual diastereomers (as described above) on
neuroprotection, 7-O-feruloylsilybin A (11a) and 7-
O-feruloylsilybin B (11b) were synthesized from the
respective pure diastereomers (Table 3).
The reaction of ferulic acid (2) and silibinin (1)
represents the first case of selective 7-position
esterification under Mitsunobu conditions.
It is noteworthy that a similar derivative of 6a, 23-O-
(4-O-acetyl)feruloylsilibinin was claimed to have
been synthesized by Liu et al. [35]. The same reaction
was repeated in our lab, but the obtained compound
The only side products observed in NMR-
measurements of the crude products are di-esters.
Integrals of H-NMR indicated that to some extent
1
two acid moieties reacted with silibinin (1); which is
Table 2. NMR data of free OH groups of silibinin derivatives.
23-O-[3,4-O-(1-
Methylethylidene)caffeoyl]-
silibinin (4a)
3, 5, 7, 20-
Tetraacetylsilibinin
(4d) [4]
7-O-(4-O-
7-O-Feruloylsilibinin
(6a)
7-O-Acetylsilibinin
(4b)
23-O-Acetylsilibinin
(4c) [31, 34]
OH groups
Acetylferuloyl)-
silibinin (4e)
3-OH
5-OH
3.468 d (1.7)
11.078 s
3.540 s
11.199 s
11.197 s
ND
3.468 s
11.060 s
11.056 s
-
3.539 s
11.194 s
11.189 s
6.388 s
5.747 s
-
-
3.682 s
11.132 s
11.126 s
-
-
11.074 s
-
7-OH
20-OH
23-OH
-
-
5.721 s
5.727 s
-
5.738 s
1.985 s
-
5.932 s
2.370 s
1.955 dd (8.1, 5.0)
2.205 – 2.133 m
was unambiguously confirmed to be 7-O-(4-O-
not a surprise running these reactions without

ACCEPTED MANUSCRIPT
protecting groups. NMR also suggested that the
second acid

ACCEPTED MANUSCRIPT
Table 3. Reaction conditions and structures of synthesized esters (5-11)
No.
5a
Name
Structure
Yield/%
7-O-Cinnamoylsilibinin
72
5b
5c
6a
6b
6c
7-O-Dihydrocinnamoylsilibinin
7-O-Benzoylsilibinin
40
55
55
18
20
7-O-Feruloylsilibinina
7-O-Dihydroferuloylsilibinin
7-O-Vanillylsilibinin
7-O-Sinapoylsilibinin
11
50
7a
7b
7-O-Dihydrosinapoylsilibinin
7-O-Syringoylsilibinin
25
7c
7-O-[3,4-O-(1-Methylethylidene)caffeoyl]silibinin
7-O-[3,4-O-(1-Methylethylidene)dihydrocaffeoyl]silibinin
7-O-[3,4-O-(1-Methylethylidene) 3,4-dihydroxybenzyl]silibinin
23-O-Cinnamoylsilibininb
63
40
37
37
8a
8b
8c
9
7-O-[4-((4-Hydroxy-3,5-dimethoxybenzoyl)-oxy)-3,5-
7
10
dimethoxybenzoyl]silibininc
7-O-Feruloylsilybin Ad
7-O-Feruloylsilybin Be
23
27
11a
11b
a: pyridine was used as base and solvent. b: 1 equivalent of NEt3 was used in the reaction. In this case the general structure does not apply as the 23-O- ester
of silibinin was yielded. c: byproduct in the synthesis of compound 7c. d: silybin A (1a) was used as starting material. e: silybin B (1b) was used as starting
material.

ACCEPTED MANUSCRIPT
moiety was not connected to another silibinin OH-
silibinin (Table 3) by regioselective esterification
with the respective acyl chlorides in moderate to good
yields without the use of protecting groups.
group. In the conversion of syringic acid into its
respective ester 7c we were able to isolate one of the
di-esters. The isolated byproduct proved to be
compound 10 (Table 3) and therefore the second acid
moiety was connected to the first through a second
ester bond. It is reasonable to assume that the main
byproducts in our reactions were the same kind of di-
2.2 Biological activities
2.2.1 Antioxidant properties (FRAP assay)
1
For the determination of the physicochemical
antioxidant properties of the compounds the ferric
reducing antioxidant power (FRAP) assay was chosen
[37, 38]. In this assay an unspecific redox-reaction
between the antioxidant (reductant) and an excess of
the ferric-tripyridyltriazine (FeIII-TPTZ) complex
takes place, producing the ferrous form (FeII-TPTZ).
The ferrous complex shows an intense blue color with
an absorption maximum of 593 nm. Therefore, with a
reducing agent present, changes in absorption can be
determined and quantified. Iron(II)sulfate is used for
calibration and as reference. Results are expressed as
equivalents of 1 mM FeSO4 reducing FeIII-TPTZ to
FeII-TPTZ. Compounds 5-10, as well as ascorbic acid,
trolox and silibinin (1) for comparison, were tested
(Figure 4).
esters as in the above case, because H-NMRs of the
crude products did not show any shifts expected for
3-O-, 20-O- and 23-O-esters.
It was also attempted to synthesize 7-O-
caffeoylsilibinin. The usual acyl chloride strategy
failed in this case. The reaction progression is likely
following the path described above with byproducts
forming multi-esters of silibinin (1) and caffeic acid
itself, giving a mixture of products. The application
of the described acyl chloride method therefore hits
its limit when using aromatic acids bearing more than
one free OH-group as substituents. As this attempt
failed, some protecting groups for the synthesis of 7-
O-caffeoylsilibinin were investigated. An acetonide
protecting group (compound 8a), benzyl protection
and silyl ethers were tested. All attempts to deprotect
the esters failed, as the compounds either showed no
conversion or decomposition upon harshening the
deprotection conditions.
Ascorbic acid and trolox showed 2 and 1.8 FeSO4
equivalents respectively, whereas silibinin (1) reached
0.6 equivalents. Compared to silibinin (1),
compounds 5 and 8 lacking a hydroxyl group in their
In summary, initial tests utilizing Mitsunobu
conditions gave either 23-O- or 7-O- silibinin esters
depending on the nature of the acid applied. This
method was therefore not suitable for our application
due to its unpredictable outcome. However, we were
able to synthesize altogether fourteen 7-O-esters of
acid-part, showed
a
significant decrease in
(physicochemical) antioxidant capacity. With
compounds 6 and 7, which bear an OH-group in their
acid-part, the reducing ability of silibinin (1) was
more or less maintained. These results show that by
esterification and therefore blocking one OH-group of
Figure 4. Compounds (1 and 5-10) were studied in the FRAP-assay. FeSO4 was used for calibration. Results are presented as
equivalents of 1 mM FeSO4 (n = 3-5; mean ± SD). Significance was defined as *p < 0.05.

ACCEPTED MANUSCRIPT
the dihydro-acids (compounds b) and the benzoic
acid ones (compounds c in Table 3 and Figure 4): the
benzoic acid compounds (c) are the least active ones,
the Michael-system compounds (a) showed higher
activities and the dihydro-compounds (b) were the
most potent antioxidants in the FRAP-assay. It has
been reported before that – albeit extensively applied
in the literature – the biophysical data as determined
in ORAC or FRAP assays gives only rough
estimations of the actual activity of the compounds as
compared with more elaborate cell-based assays and
often the results are even diverge [13, 39, 40].
Figure 5. Compounds (6a, 11a, 11b and 5a) were studied in
comparison to silibinin (1), ferulic acid (2), cinnamic acid (CA)
and the respective 1:1 mixtures. FeSO4 was used for
calibration. Results are presented as equivalents of 1 mM
FeSO4 (n = 3-5; mean ± SD). Significance was defined as *p <
2.2.2 Neuroprotective effects
For the evaluation of the neuroprotective effects the
HT-22 hippocampal neuronal cell-model was used.
This neuronal cell line is derived from murine
hippocampal tissue and is glutamate sensitive [41].
As these cells lack ionotropic glutamate receptors,
addition of extracellular glutamate in high
concentrations inhibits cystine (the oxidized form of
cysteine) transport via the cystine/glutamate
antiporter. This leads to intracellular cysteine and
therefore glutathione depletion which induces
intracellular ROS accumulation and cell injuries
resulting in neuronal oxidative stress and cell death
[42-44].
silibinin (1) the antioxidant effect was reduced if
there was no additional OH-group present in the
molecule. Furthermore, it can be observed that the
respective dihydro-derivatives (compounds b)
showed the most potent antioxidant properties
compared to their unsaturated counterparts
(compounds a) and the compounds lacking the C2
moiety (compounds c). 7-O-Dihydrosinapoylsilibinin
(7b) was the most potent antioxidant, reaching an
effect similar to trolox.
7-O-Feruloylsilibin 6a, its diastereomers 11a and
11b, as well as 7-O-cinnamoylsilibinin 5a were
compared to silibinin (1) and the individual acids
(cinnamic acid and ferulic acid (2)) and equimolar
mixtures thereof (Figure 5).
We first screened all the synthesized compounds (5-
11) at concentrations of 10 µM and 25 µM for their
neuroprotective potential in the glutamate assay.
Figure 6 shows the cell viability as percentage of the
control for the respective compounds. Quercetin at 25
µM served as a positive control with about 80% of
the cells surviving glutamate treatment. Comparing
compounds 5-8, the most potent neuroprotectants
were the compounds carrying the Michael-system
(5a, 6a, 7a, 8a), whereby the respective dihydro-
derivatives only showed weak (6b, 7b) or no
neuroprotection (5b, 8b). The biological activities of
the benzyl-derivatives (5c, 6c, 7c) were very
dependent on their aromatic substituents and started
exhibiting activity only at 25 µM. The benzyl ester 5c
displayed the weakest activity, whereby the vanillyl
ester 6c and syringoyl ester 7c both showed good or
moderate activity, respectively. The highest
neuroprotective activities were shown by 7-O-
cinnamoylsilibinin (5a), 7-O-feruloylsilibinin (6a),
the protected caffeic acid ester 8a and, surprisingly,
the protected dihydroxybenzoic acid ester 8c already
at 10 µM. At a concentration of 25 µM their
neuroprotective effects declined, which is most likely
Ferulic acid (2) showed activity similar to trolox and
the 1:1 mixture of silibinin (1) and ferulic acid (2)
displayed higher antioxidative properties albeit not
statistically significantly different to ferulic acid (2)
alone. In contrast, the corresponding esters 6a, 11a
and 11b showed lower activity than silibinin (1). CA
did not reveal any activity in the FRAP-assay (data
not shown) and the equimolar mixture of silibinin (1)
and CA showed the same effect as silibinin (1). No
difference between the silibinin-ferulic ester
diastereomers 11a and 11b was observed.
Taken together, ferulic acid (2) was shown to be a
potent compound in contrast to CA. No statistically
significant overadditive effect of the equimolar
mixtures over silibinin (1) could be observed and
esterification of silibinin generally had a negative
effect on antioxidant properties measured in vitro in
the physicochemical FRAP-assay. Only 7-O-
dihydrosinapoylsilibinin (7b) showed improved
activity. There are surprisingly clear cut SARs when
comparing the α,β-unsaturated (compounds a) with

ACCEPTED MANUSCRIPT
120
100
80
60
40
20
0
***
***
***
***
***
***
***
***
***
***
***
***
***
***
**
***
***
*
***
**
*
**
Glutamate 5 mM - + + + + + + + +
Quercetin 25 µM - - + - - - - - -
Compound [µM] - - -
+ + + + + + + + + + + + + + + + + + + +
+ + + + + +
- - - - - -
- - - - - - - - - - - -
- -
- -
- - - -
5b 5c
9
11a 11b
5a
6b 6c
7b 7c
8b 8c
8a
10
6a
7a
Figure 6. Compounds (5-11) were studied on neuronal HT-22 cells for neuroprotection against glutamate induced oxidative stress at 10 µM and
25 µM. Results of the modified MTT test are presented as means ± SD and refer to untreated control cells which were set as 100% values.
Levels of significance: * p < 0.05; ** p < 0.01; *** p < 0.001.
due to innate cytotoxic effects as described before for
other antioxidants, e. g. heterocycles (cf. below) [18].
Furthermore, 23-O-Cinnamoylsilibinin (9) only
shows moderate neuroprotection at 25 µM and none
at 10 µM. This result compared to that of the 7-O-
ester 5a shows that blocking of the 7-OH group of
silibinin has a positive influence on neuroprotection.
The conjugation of two moieties (in this case
cinnamic acid and silibinin) at the 7-O- position is
therefore superior to 23-esterfication, which supports
the intended aim of our study.
25 µM, its extent being comparable to the positive
control quercetin. This clearly demonstrates the
overadditive effect over the individual components
and their equimolar mixtures.
A
120
*
*
100
80
60
40
20
0
*
***
Two of the most potent compounds, 5a and 6a, were
further examined by investigating additional
concentrations (1, 5, 10 and 25 µM) and comparing
their neuroprotective features to the respective
individual compounds as well as the 1:1 (equimolar)
mixtures thereof. Furthermore, neurotoxicity was
determined in an additional MTT-assay.
[µM]
1
5
10 25
1
5
10 25
1
5
10 25
1
5
10 25
1
CA + 1
5a
CA
B
120
100
80
60
40
20
0
***
***
***
***
Potential
neurotoxic
effects
of
7-O-
cinnamoylsilibinin (5a) in comparison to silibinin (1),
CA and the 1:1 mixture of CA and silibinin (1) were
evaluated (Figure 7A). Only at the highest
concentration of 25 µM, both the 1:1 mixture and
Glutamate 5 mM
Quercetin 25 µM
Compound [µM]
-
-
-
+
-
-
+
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
1
5
10 25
CA
1
5
10 25
1
5
10 25
1
5
10 25
5a
compound
5a
showed
slightly
increased
CA + 1
1
neurotoxicity. This effect explains the decreased
neuroprotection at 25 µM observed in the screening
performed at 10 and 25 µM (Figure 6). Figure 7B
shows the neuroprotective effects of the
aforementioned compounds. Again, only compound
5a showed neuroprotection at 5 µM, 10 µM and
Figure 7. Compounds 1 and 5a, as well as cinnamic acid (CA) and
the 1:1 mixture of 1 and CA, were studied on neuronal HT-22 cells
for A) neurotoxic effects and B) neuroprotection against glutamate
induced oxidative stress at 1-25 µM. Results of the modified MTT
test are presented as means ± SD of three different independent
experiments and refer to untreated control cells which were set as
100% values. Levels of significance: * p < 0.05; ** p < 0.01; *** p
< 0.001.

ACCEPTED MANUSCRIPT
In Figure 8A the neurotoxic effects of 7-O-
be investigated in future work.
feruloylsilibinin (6a) and its diastereomers 11a and
11b compared to silibinin (1), ferulic acid (2) and the
1:1 mixture of the components are depicted. Silibinin
(1) and ferulic acid (2) exhibited no neurotoxicity,
while the 1:1 mixture and the ester 6a showed
moderate toxicity at higher concentrations. The
neuroprotection effects are shown in Figure 8B.
Compound 6a displayed neuroprotection at 10 µM
and 25 µM with cell survival of about 50%. The
individual components lacked any neuroprotection
and furthermore the 1:1 mixture of silibinin (1) and
ferulic acid (2) exposed no neuroprotection,
underlining the overadditive effect of 7-O-
feruloylsilibinin (6a).
In summary, the synthesized compounds 5-11 were
screened for their neuroprotective effects on HT-22
cells and compounds bearing a Michael system (5a,
6a, 8a) proved to be the most potent. Concentration-
dependent evaluation of the biological activity of
compounds 5a and 6a in comparison to their
individual components and the respective equimolar
mixtures
revealed
that
the
esters
7-O-
cinnamoylsilibinin (5a) and 7-O-feruloylsilibinin (6a)
show a very pronounced overadditive neuroprotective
effect. Furthermore, the diastereomer 7-O-
feruloylsilybin B (11b)
displays
higher
neuroprotection at 10 µM than its counterpart 11a.
Subsequently, the esters 7-O-cinnamoylsilibinin (5a)
and 7-O-feruloylsilibinin (6a) including the pair of
diastereomeric esters (11a and 11b) were tested in a
portfolio of other biological assays with regard to
neuroprotection, namely in vitro ischemia, inhibition
of microglial activation and a PC12 differentiation
assay [27].
In addition, the pure diastereomers of the ferulic acid
ester 11a and 11b were tested (Figure 8). In terms of
neurotoxicity at 10 µM 11a exhibited a slightly
higher neurotoxic effect than 11b at 10 µM. This is in
accordance with the increased neuroprotection of 11b
at this concentration (Figure 8B). More remarkable
and in accordance with reports in the literature [19,
21-25], the esters of the diastereomers showed a
statistically significant difference in the extent of
neuroprotection at 10 µM, when 11b was more
potent. The – albeit not pronounced – but significant
difference in biological activity of two diastereomeric
esters indicates a specific mode of action that has to
In the in vitro ischemia model in HT-22 cells a loss of
adenosine triphosphate (ATP), which is associated
with neuronal cell damage and death in AD [45], in
HT-22 cells is induced by addition of iodoacetic acid
and the cell viability is determined by an MTT test.
Inhibition of activated pro-inflammatory microglia,
which produce a variety of cytotoxic and pro-
inflammatory factors promoting neurodegeneration,
may have a positive effect on AD pathology [46]. In
this assay mouse BV2 microglial cells are treated
with bacterial lipopolysaccharide in the presence of
the compound and the resulting nitrite concentration
in the culture medium is quantified.
A
120
***
**
***
100
80
60
40
20
0
***
***
***
***
25
Regarding neurite outgrowth, the connection of
nerves in AD is greatly impaired and potential
neuroregeneration would be a particular benefit. A
well-studied model uses neurite outgrowth of rat
PC12 cells after nerve growth factor (NGF)
stimulation. After treatment with the respective
compounds, cells that developed neurites are counted
[47].
1
5
10 25
6a
1
5
10
[µM]
1
5
10 25
1
1
5
10 25
2
1
5
10 25
1
5
10 25
1+2
11a
11b
B
120
100
80
60
40
20
0
$$$
***
***
***
***
***
***
As shown in Table 4, neither ferulic acid (2) nor
cinnamic acid (CA) and silibinin (1) showed
significant activity in any of the described assays. The
latter was moderately active in the PC12
differentiation assay only, with no activity in the
microglia activation and ischemia model. In contrast
to this, 7-O-cinnamoylsilibinin (5a) and both
diastereomers of 7-O-feruloylsilibinin (11a and 11b)
Glutamate 5 mM
Quercetin 25 µM
Compound [µM]
-
-
-
+
-
-
+
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
1
5 10 25
1
5 10 25
1
5 10 25
1
5 10 25
1
5 10 25
1
5 10 25
11b
1
2
1+2
6a
11a
Figure 8. Compounds 1, 2, 6a, 11a and 11b, as well as the 1:1 mixture
of 1 and 2, were studied on neuronal HT-22 cells for A) neurotoxic
effects and B) neuroprotection against glutamate induced oxidative
stress at 1-25 µM. Results of the modified MTT test are presented as
means ± SD of three different independent experiments and refer to
untreated control cells which were set as 100% values. Levels of
significance: * p < 0.05; ** p < 0.01; *** p < 0.001. Levels of
significance for Bonferroni’s posttest are indicated as $.

ACCEPTED MANUSCRIPT
showed low micromolar activity in the iodoacetic
formation of these aggregates may therefore be a
viable option in the treatment of AD.
acid-induced ATP depletion of HT-22 cells (Table 4).
Regarding the inhibition of microglia activation with
concomitant production of pro-inflammatory and
cytotoxic factors, the compounds showed activity
ranging from 4.2 to 10.1 µM, with 7-O-
cinnamoylsilibinin (5a) being the most active one.
Again, none of the components (i. e. silibinin or the
acids) showed any activity. Investigating PC12
differentiation, silibinin (1) showed 38% neurite
outgrowth, 7-O-cinnamoylsilibinin (5a) 40% and
both diastereomers of 7-O-feruloylsilibinin (11a and
11b) between 50 and 67%. As for the in vitro
ischemia assay, the silybin B-based diastereomer
(11b) seems to be more active, but statistical
A screening method for anti-aggregation capacities
was recently published [28, 29]. This assay uses
bacterial cells overexpressing Aβ42 or τ protein.
Inclusion bodies (IBs), possessing amyloid-like
properties, are formed by Aβ in bacteria. These IBs
are stained by thioflavin-S (Th-S) to assess the
amount of Aβ aggregated.
Table 5 shows the inhibition effects against both
Aβ42 peptide and τ protein of compounds 5a, 6a, 11a
and 11b at 10 µM and 100 µM. In all cases the very
modest effects of silibinin (1) were greatly exceeded
by the test compounds. Regarding Aβ42 aggregation
the esters showed a two- to three-fold higher
inhibition than silibinin (1), with the cinnamic- (5a)
and ferulic acid esters (6a) being the most potent. For
the inhibition of τ protein aggregation the effects are
even more pronounced. At only 10 µM the esters
show at least a 15-fold higher inhibition than silibinin
(1) and an overall inhibition of 30% (Table 5).
Table 4. Biological activities of selected compounds in following
assays: in vitro ischemia, inhibition of microglial activation and
ability to promote differentiation of PC12 cells
Inhibition of
in vitro
ischemia EC50
(µM)*
PC12 cells
differentiation
(at 10 µM)
microglial
activation
EC50 (µM)*
No.
A significant difference in effects of the individual
diastereomers 11a and 11b can be observed in both
assays. The half-maximal inhibitory concentrations
(IC50) were determined for both compounds.
Diastereomer 11a showed IC50 values of 85.5 ± 5.8
µM against Aβ42 aggregation and 27.2 ± 2.3 µM
against τ protein aggregation. 7-O-Feruloylsilybin B
(11b) showed IC50 values of 43.6 ± 4.8 µM against
Aβ42 aggregation and 12.9 ± 1.2 µM against τ
protein aggregation. Thus, diastereomer 11b is more
potent than its counterpart 11a in the aggregation
>10
>10
>10
38 ± 3%
no
1
>10
2
CA
5a
>10
>10
no
1.5 ± 0.2
>10
4.2 ± 0.9
6.5 ± 0.8
9.8 ± 0.9
10.1 ± 3.3
40 ± 20%
47 ± 12%
50 ± 10%
67 ± 12%
6a
1.4 ± 0.5
2.2 ± 0.05
11a
11b
*EC50 (µM) – half-maximal effective concentration
significance was not reached.
Taken together, three additional assays regarding
neuroprotection and other activities related to AD
showed an overall lack of activity for the natural
carboxylic acids (cinnamic and ferulic acid) and
silibinin (1), despite the latter showing some activity
on PC12 differentiation. 7-O-Cinnamoylsilibinin (5a)
as well as both diastereomers of 7-O-feruloylsilibinin
(11a and 11b) showed good activities in all three
assays with 7-O-cinnamoylsilibinin (5a) being the
most active one on inhibition of microglial activation.
Table 5. Effect of anti-aggregating compounds monitored by Th-S
staining of bacterial cells overexpressing Aβ42 peptide or τ protein
Aβ42 aggregation
inhibition
τ protein aggregation
inhibition
No.
(% ± SEM)
(% ± SEM)
10 µM
100 µM
10 µM
100 µM
0.0 ± 1.0
0.0 ± 4.4
0.0 ± 1.4
0.0 ± 2.4
Control
1
4.8 ± 3.1
16.0 ± 4.2
61.1 ± 4.6
54.4 ± 3.3
52.7 ± 3.2
73.0 ± 3.7
2.0 ± 2.7
37.1 ± 5.5
30.4 ± 3.2
36.8 ± 4.6
44.5 ± 2.2
10.1 ± 3.8
76.1 ± 1.7
64.7 ± 1.6
67.2 ± 4.9
84.0 ± 1.0
17.1 ± 2.4
16.8 ± 1.7
13.2 ± 2.3
14.2 ± 3.8
5a
6a
11a
11b
2.2.3 Inhibition of Aβ42 and τ protein aggregation
Additional hallmarks of AD are the formation of
extracellular amyloid plaques and neurofibrillary
tangles, formed by hyperphosphorylated τ protein.
These deposits are toxic and lead to neuronal cell
death [48, 49]. Compounds that are able to inhibit the
inhibition assays.
2.3 Stability of 7-O-feruloylsilibinin (6a)

ACCEPTED MANUSCRIPT
The stability of 7-O-feruloylsilibinin (6a) as a
silibinin (1) and ferulic acid (2).
representative compound was investigated to ensure
the observed effects derive from the compounds
themselves and not from the mixture of its
components after hydrolysis (which has in fact been
ruled out by comparison with the equimolar mixture)
or other products formed. This was accomplished by
HPLC experiments. 7-O-Feruloylsilibinin (6a) was
dissolved in assay medium and human pooled
plasma, respectively, and incubated for 24 h. Samples
were taken after certain times and the area under
curve was determined.
Figure 9B shows the time-dependent stability of 6a in
human pooled plasma. Conversion is, as expected,
more rapid than in medium with an approximate half-
life of 1.5 hours and roughly 20% of the compound
remaining after four hours. In plasma, the formation
of
7-O-feruloyl-2,3-dehydrosilibinin
was
not
detected, but rather the cleavage of the ester bond
takes place.
3. Conclusion
Figure 9A shows the time-dependent stability of 6a in
assay medium. The graph shows a rather fast
conversion of 6a during the initial hours but even
after 24 h 25% of 6a is still present. LCMS
investigations indicated that the disappearance of 6a
is not due to hydrolysis of the ester bond, but rather
formation of the respective 2,3-dehydrosilibinin ester
in assay medium. This is in accordance with previous
reports on the stability of silibinin esters and the
assumption that hydrolysis is reduced by inhibition of
esterases [17]. This of course also explains the
pronounced neuroprotection that is absent for both the
individual components as well as the mixture of
Conjugates of the flavonolignan silibinin (1) and
derivatives of ferulic acid (2) were synthesized by
esterification via the respective acyl chlorides. This
method avoids the use of protective groups and
enables selective esterification at the 7-position of
silibinin (1). Only 7-O-dihydrosinapoylsilibinin (7b)
showed improved antioxidant capacity compared to
silibinin (1) in the cell-free FRAP-assay. Despite their
lower physicochemical antioxidant properties the
esters
7-O-cinnamoylsilibinin
(5a),
7-O-
feruloylsilibinin (6a) and two acetonide-protected
caffeoyl esters 8a and 8c showed pronounced
neuroprotection
regarding
glutamate-induced
oxidative stress-dependent cell death in murine
hippocampal cells at 10 µM. These neuroprotective
effects were absent for the individual components and
A
mixtures
thereof,
which
demonstrates
the
100
overadditive neuroprotection of the prepared esters.
This pronounced overadditive effect was also
observed in inhibition of microglia activation,
neuroregeneration measured by neurite outgrowth on
PC12 cells, and in an in vitro ischemia model. Since
several distinct assays indicated a difference between
the two diastereomers of 7-O-feruloylsilibinin (6a),
we investigated their ability to inhibit aggregation of
Aβ42 peptide and τ protein. Again, all esters tested
were much more potent than the components and in
both aggregation assays a statistically significant
difference between the diastereomers was observed.
Stability measurements in cell medium revealed that
the compounds are gradually oxidized (or
disproportionated [50]) to the respective 2,3-
dehydrosilibinin derivatives before hydrolysis of the
ester bond takes place, supporting the assumption that
the esters are the bioactive compounds.
75
50
25
0
0
1
2
3
4
5
23 24 25
time [h]
B
100
75
50
25
0
The esters synthesized are comparatively stable
towards hydrolysis, even in plasma. In fact the
formation of esters can be used to protect the pro-
oxidative and metabolically labile 7-OH group of
silibinin (1). We have shown that 7-O-esters of
0
2
4
6
8
10 23 24 25
time [h]
Figure 9. Time-dependent analysis of the stability of compound
6a in A) assay medium and B) human pooled plasma (n = 6).
AUC was determined by HPLC. AUC at 0 min was set as 100%.
The experiments were performed in triplicate (mean ± SD).

ACCEPTED MANUSCRIPT
acquired on a Shimadzu LCMS 2020. Analytic HPLC was
performed on a system from Shimadzu Products equipped
with DGU-20A3R controller, LC20AB liquid
chromatograph, and SPD-20A UV/Vis detector.
Stationary phase was a Synergi 4U fusion-RP (150 × 4.6
mm) column. Mobile phase: gradient MeOH
(0.01%TFA)/water (0.01%TFA) (phase A/ phase B) were
used. For analytical HPLC a flow rate of 1 mL/min and for
preparative HPLC a flow rate of 3 mL/min was used.
silibinin with natural antioxidant carboxylic acids
show pronounced neuroprotective effects exceeding
the ones of the components. It is promising that the
cinnamic acid esters that lack hydroxy groups at the
benzene ring are so potent, since they are less prone
to rapid phase II metabolism and in the literature, in
most cases, caffeic and ferulic acid derivatives have
been described as antioxidant components in hybrid
molecules [2 and numerous references therein].
Remarkably, there are differences between the
silibinin-diastereomeric esters, which might indicate
specific and targeted modes of action that need to be
further elucidated in future work. As reported for
a
a
General Procedure for the synthesis of silibinin esters
(5-11):
To a solution or suspension of the respective acid (0.315
mmol, 1.05 eq) in 5 mL of dichloromethane, 2 µL of
dimethylformamide and 28.4 µL of oxalylchloride (0.33
mmol, 1.1 eq) were added. The resulting mixture was
stirred at room temperature for 1 h and was added to a
solution of 146 mg of silibinin (0.3 mmol, 1 eq) and 192 µL
of triethylamine (1.35 mmol, 4.5 eq) in 10 mL
tetrahydrofuran. The reaction mixture was stirred at room
temperature for 2-16 h. The reaction was quenched with 30
mL of water and extracted with ethyl acetate (3x 30 mL).
The combined organic phases were washed with 10 mL of
1M HCl-solution and 20 mL of brine, dried over sodium
sulfate and the solvent was removed in vacuo.
silybin
A and B [22], we could observe a
stereospecific effect on amyloid aggregation, as well
as τ protein aggregation, of the esters tested, but much
more pronounced than for the silibinins.
The most potent compounds in the cell-based assays
contain a Michael system. Due to the presence of a
Michael system in the molecules, it is important to
mention that the assays applied to estimate biological
activity have to be carefully chosen (“PAINS”) to
avoid false-positive readouts, which is not the case
for both the FRAP and the cell-based assays
described and used herein [51]. Since in the long term
the ester compounds are metabolized to known non-
toxic compounds, such esters may represent a new
approach for anti-neuroinflammatory approaches in
medicinal chemistry.
7-O-Cinnamoylsilibinin (5a) Following the general
procedure 47 mg of cinnamic acid (0.315 mmol, 1.05 eq)
were used. Column chromatography (CH2Cl2 : MeOH =
50:1) gave 133 mg (0.211 mmol, 72 %) of the desired
1
product as a white solid. HPLC-purity: 98.4 % H-NMR
(400 MHz, CDCl3) δ (ppm) = 2.12 (br. s., 1H, 23-OH),
3.56 (dd, J = 12.4, 3.6 Hz, 1 H, 23a-H), 3.80 (dd, J = 12.4,
2.6 Hz, 1 H, 23b-H), 3.91 (m, 3 H, OMe), 4.04 (dddd, J =
8.3, 4.0, 2.6, 1.4 Hz, 1 H, 10-H), 4.60 (dd, J = 12.3, 5.5 Hz,
1 H, 3-H), 4.95 (d, J = 8.3 Hz, 1 H, 11-H), 5.06 (d, J = 12.3
Hz, 1 H, 2-H), 5.91 (br. s., 1H, 20-OH), 6.40 (dd, J = 5.3,
2.0 Hz, 1 H, 8-H), 6.46 (d, J = 2.0 Hz, 1 H, 6-H), 6.57 (dd,
J = 15.8, 1.3 Hz, 1 H, 2’-H), 6.92 – 6.96 (m, 3 H, 18-H, 21-
H, 22-H), 7.02 – 7.11 (m, 2 H, 15-H, 16-H), 7.19 (dd, J =
6.8, 2.0 Hz, 1 H, 13-H), 7.39 – 7.46 (m, 3 H, b-H, d-H, f-
H), 7.55 – 7.61 (m, 2 H, c-H, e-H), 7.86 (dd, J = 16.1, 0.5
Hz, 1 H, 3’-H), 11.10 (br. s, 1 H, 5-OH). 13C-NMR
(100 MHz, CDCl3) δ (ppm) = 60.6 (23-CH2), 55.1 (OMe),
71.60/71.68 (3-CH), 75.34/75.38 (11-CH), 77.32/77.36 (10-
CH), 82.12/82.16 (2-CH), 101.2 (8-CH), 102.6 (6-CH),
103.1 (4a-Cq), 108.53/108.57 (18-CH), 113.7 (21-CH),
115.4 (2’-CH), 115.52/115.62 (13-CH), 116.24/116.34 (16-
CH), 119.80 (22-CH), 119.93/120.14 (15-CH), 126.8 (17-
Cq), 127.4 (c-CH, e-CH), 128.05/128.08 (14-Cq), 128.1 (b-
CH, f-CH), 130.1 (d-CH), 132.9 (a-Cq), 142.9 (12a-Cq),
143.2 (16-Cq), 145.4 (20-Cq), 146.0 (19-Cq), 146.8 (3’-CH),
158.4 (8a-Cq), 161.30/161.32 (7a-Cq), 161.8 (5-OH), 163.1
(1’-COO), 196.4 (4-CO). ESI-MS: 613.20 m/z [M+H]+,
635.60 m/z [M+Na]+.
In future work the influence of the 2,3-dehydro
derivatives must be determined, since the activity of
these conversion products is not elucidated yet.
Furthermore, the ability of the compounds to cross
the blood brain barrier as well as their lipophilic
parameters and solubility need to be quantified.
4. Experimental
4.1 Chemical synthesis
General
Tetrahydrofuran (THF) was freshly distilled from
sodium/benzophenone. Dichloromethane was dried over
calcium hydride and distilled. Common reagents and
solvents were from commercial suppliers and used as such.
Reaction progress was monitored using analytical thin layer
chromatography (TLC) on precoated silica gel GF254
plates (Macherey-Nagel GmbH
& Co. KG, Düren,
Germany), the spots were detected under UV light (254
nm). Column chromatography was conducted using silica
gel (high-purity grade, 0.035-0.070 mm, Merck KGaA,
Darmstadt, Germany). Nuclear magnetic resonance spectra
were performed with Bruker AV-400 NMR instrument
(Bruker, Karlsruhe, Germany). ESI mass spectral data was
7-O-Dihydrocinnamoylsilibinin (5b) Following the
general procedure, 47 mg of dihydro cinnamic acid (0.315
mmol, 1.05 eq) were used. Column chromatography

ACCEPTED MANUSCRIPT
(CH2Cl2 : MeOH = 30:1) followed by preparative HPLC
(8a-Cq),
162.94/162.95
(5-OH),
164.1
(COO),
(Method C) gave 74 mg (0.12 mmol, 40 %) of the desired
197.56/197.57 (4-CO). ESI-MS: 585.10 m/z [M-H]-,
609.15 m/z [M+Na]+.
1
product as a white solid. HPLC-purity: 98.6 % H-NMR
(400 MHz, CDCl3) δ (ppm) = 2.87 (td, J = 7.8, 0.5 Hz, 2 H,
2’-CH2), 3.05 (t, J = 7.5 Hz, 2 H, 3’-CH2), 3.55 (dd, J =
12.4, 3.9 Hz, 1 H, 23a-H), 3.80 (dd, J = 12.4, 2.6 Hz, 1 H,
23b-H), 3.90 (s, 3 H, OMe), 4.04 (ddd, J = 8.0, 3.8, 2.5 Hz,
1 H, 10-H), 4.56 (dd, J = 12.2, 4.6 Hz, 1 H, 3-H), 4.93 (dd,
J = 8.3, 1.0 Hz, 1 H, 11-H), 5.02 (d, J = 12.0 Hz, 1 H, 2-H),
6.24 (dd, J = 5.5, 2.0 Hz, 1 H, 8-H), 6.29 (dd, J = 2.0, 0.8
Hz, 1 H, 6-H), 6.91 – 6.95 (m, 3 H, 18-H, 21-H, 22-H), 7.02
– 7.07 (m, 2 H. 15-H, 16-H), 7.17 (dd, J = 6.1, 1.9 Hz, 1 H,
13-H), 7.22 – 7.25 (m, 3 H, b-H, d-H, f-H), 7.29 – 7.35 (m,
2 H, c-H, e-H), 11.08 (d, J = 2.5 Hz, 1 H, 5-OH). 13C-NMR
(100 MHz, CDCl3) δ (ppm) = 30.8 (3’-CH2), 36.0 (2’-
CH2), 56.1 (OMe), 61.7 (23-CH2), 72.61/72.67 (3-CH),
76.35/76.39 (11-CH), 78.35/78.39 (10-CH), 83.12/83.15 (2-
CH), 102.1 (8-CH), 103.6 (6-CH), 104.1 (4a-C),
109.62/109.65 (18-CH), 114.8 (21-CH), 116.55/116.64 (13-
7-O-Feruloylsilibinin (6a) Mitsunobu reaction: To a
solution of 145 mg of silibinin (0.3 mmol, 1 eq) in THF (15
mL) 116 mg of ferulic acid (0.6 mmol, 2.0 eq) and 197 mg
of triphenylphosphine (0.75 mmol, 2.5 eq) were added. A
solution of 147 µL of diisopropyl azodicarboxylate (0.76
mmol, 2.5 eq) in THF (7 mL) was added dropwise at 10°C
(water bath) over 4 h. The speed of addition was kept slow
so that no solid precipitated in the reaction mixture. After
the addition, the mixture was stirred at room temperature
for 20 h. The reaction was quenched with water, and
extracted with ethyl acetate (3 × 30 mL). The combined
organic phases were washed with brine, dried over sodium
sulfate and the solvent was removed in vacuo. The residue
was purified by column chromatography (CH2Cl2 : MeOH
= 50:1) twice to give 89 mg (0.13 mmol, 45 %) of the
desired product as a yellow foam. HPLC-purity: 97.9 %
1H-NMR (400 MHz, CDCl3) δ(ppm) = 1.96 (dd, J = 8.1,
5.0 Hz, 1H, 23-OH), 3.49 (d, J = 1.7 Hz, 1H, 3-OH), 3.57
(ddd, J = 12.3, 8.1, 4.0 Hz, 1H, 23a-H), 3.81 (ddd, J = 12.5,
5.0, 2.8 Hz, 1H, 23b-H), 3.92 (s, 3H, OMe), 3.95 (s, 3H,
OMe), 4.02 – 4.09 (m, 1H, 10-H), 4.60/4.61 (dd, J = 12.1,
1.7 Hz, 1H, 3-H), 4.97 (d, J = 8.3 Hz, 1H, 11-H), 5.07 (d, J
= 12.1 Hz, 1H, 2-H), 5.72 (s, 1H, 20-OH), 5.95 (s, 1H, d-
OH), 6.39/6.40 (d, J = 2.0 Hz, 1H, 8-H), 6.41/6.41 (d, J =
15.8 Hz, 1H, 2’-H), 6.45 (d, J = 2.0 Hz, 1H, 6-H), 6.92 –
6.98 (m, 4H, e-H, 18-H, 21-H, 22-H), 7.03 – 7.12 (m, 3H,
b-H, 15-H, 16-H), 7.14 (d, J = 8.3, Hz, 1H, f-H), 7.19/7.21
(d, J = 1.8 Hz, 1H, 13-H), 7.79 (d, J = 15.8 Hz, 1H, 3’-H),
11.07/11.08 (s, 1H, 5-OH). 13C-NMR (100 MHz, CDCl3)
δ(ppm) = 56.15 (OMe), 56.20/56.21 (OMe), 61.8 (23-CH2),
72.74/72.83 (3-CH), 76.50/76.54 (11-CH), 78.46/78.49 (10-
CH), 83.27/83.30 (2-CH), 102.3 (8-CH), 103.8 (6-CH),
104.1 (4a-Cq), 109.64/109.67 (18-CH), 109.8 (b-CH), 113.7
(2’-CH), 114.8 (21-CH), 115.1 (e-CH), 116.64/116.78 (13-
CH),
117.26/117.35
(16-CH),
120.8
(22-CH),
120.99/121.15 (15-CH), 126.6 (d-CH), 127.8 (17-Cq), 128.4
(b-CH, f-CH), 128.7 (c-CH, e-CH), 129.04/129.07 (14-Cq),
139.7 (a-Cq), 143.9 (12a-Cq), 144.23/144.25 (16-Cq), 146.5
(20-Cq), 147.0 (19-Cq), 159.1 (8a-Cq), 162.27/162.29 (7a-
Cq), 162.8 (5-OH), 170.2 (1’-COO), 197.5 (4-CO). ESI-
MS: 615.20 m/z [M+H]+, 637.15 m/z [M+Na]+.
7-O-Benzoylsilibinin (5c) A solution of 145 mg of silibinin
(0.30 mmol, 1 eq) in THF (15 mL) and CH2Cl2 (30 mL)
was cooled to 0°C in an ice-bath. 69 µL of triethylamine
(0.39 mmol, 1.3 eq) and 45 µL benzoyl chloride (0.39
mmol, 1.3 eq) were added dropwise. The mixture was
stirred at room temperature for 4 h. The reaction was
quenched with water, extracted with ethyl acetate (3 × 40
mL). The combined organic phases were washed with
brine, dried over sodium sulfate and the solvents removed.
Flash chromatography (CH2Cl2 : MeOH = 20:1) gave 97 mg
(0.17 mmol, 55%) of the desired product as a white
amorphous solid. HPLC-purity:
≥
99
%
1H-NMR
CH),
117.41/117.51
(16-CH),
121.0
(22-CH),
(400 MHz, CDCl3) δ (ppm) = 2.14 (s, 1H, 23-OH), 3.58 (s,
1H, 3-OH), 3.50 – 3.62 (m, 1H, 23a-H), 3.77 – 3.83 (m,
1H, 23b-H), 3.90 (s, 3H, OMe), 4.01 – 4.07 (m, 1H, 10-H),
4.59/4.61 (d, J = 12.1 Hz, 1H, 3-H), 4.94 (d, J = 8.3 Hz, 1H,
11-H), 5.06 (d, J = 12.1 Hz, 1H, 2-H), 5.81 (s, 1H, 20-OH),
6.43/6.45 (d, J = 2.0 Hz, 1H, 8-H), 6.51 (d, J = 2.0 Hz, 1H,
6-H), 6.89 – 6.98 (m, 3H, 18-H, 21-H, 22-H), 7.03/7.04 (d,
J = 8.4 Hz, 1H, 16-H), 7.08/7.09 (dd, J = 8.4, 1.9 Hz, 1H,
15-H), 7.18/7.20 (d, J = 1.8 Hz, 1H, 13-H), 7.51 (t, J = 7.6
Hz, 2H, c-H, e-H), 7.61 – 7.69 (m, 1H, d-H), 8.16 (d, J =
8.3 Hz, 2H, b-H, f-H), 11.122 (s, 1H, 5-OH). 13C-NMR
(100 MHz, CDCl3) δ (ppm) = 56.2 (OMe), 61.8 (23-CH2),
72.74/72.81 (3-CH), 76.44/76.49 (11-CH), 78.44/78.48 (10-
CH), 83.25/83.29 (2-CH), 102.4 (8-CH), 103.9 (6-CH),
104.3 (4a-Cq), 109.67/109.71 (18-CH), 114.8 (21-CH),
116.65/116.75 (13-CH), 117.36/117.46 (16-CH), 120.9 (22-
CH), 121.07/121.25 (15-CH), 127.9 (17-Cq), 128.8 (a-Cq, c-
CH, e-CH), 129.15/129.18 (14-Cq), 130.5 (b-CH, f-CH),
134.2 (d-CH), 144.0 (12a-Cq), 144.32/144.35 (16a-Cq),
146.5 (20-Cq), 147.1 (19-Cq), 159.6 (7-Cq), 162.46/162.49
121.06/121.30 (15-CH), 123.9 (f-CH), 126.6 (a-Cq), 127.9
(17-Cq), 129.20/129.25 (14-Cq), 144.08/144.10 (12a-Cq),
144.36/144.39 (16a-Cq), 146.6 (20-Cq), 147.0 (19-Cq),
147.09/147.10 (c-Cq), 148.0 (3’-CH), 148.9 (d-Cq), 159.7
(7-Cq), 162.45/162.47 (8a-Cq), 162.92/162.94 (5-OH),
164.5 (1’-COO), 197.45/197.46 (4-CO). ESI-MS: 658.50
m/z [M+H]+. Acyl chloride method: To the suspension of
35 mg of ferulic acid (0.18 mmol, 1.2 eq) in CH2Cl2 (5 mL)
was added 17 µL of oxalyl chloride (0.2 mmol, 1.3 eq) and
DMF (2 µL, catalytic amount). The mixture was stirred at
room temperature for 1 h. The resulting yellow solution was
subsequently added drop wise to another flask charged with
73 mg of silibinin (0.15 mmol, 1 eq) and pyridine (15 mL).
The reaction mixture was stirred at room temperature for 2
h, quenched with water and extracted with ethyl acetate (3
× 40 mL). The combined organic phases were washed with
1 M HCl (2 × 10 mL), and brine, dried over sodium sulfate
and the solvent was removed. Flash chromatography
(CH2Cl2 : MeOH = 50:1) gave 54 mg (0.08 mmol, 55%) of
the desired product as a light yellow foam. Analytic data

ACCEPTED MANUSCRIPT
was consistent with that of compound 6a prepared in the
above Mitsunobu reaction.
(5-OH), 163.7 (a-COO), 197.36/197.38 (4-CO). ESI-MS:
633.15 m/z [M+H]+, 655.10 m/z [M+Na]+.
7-O-Dihydroferuloylsilibinin (6b) Following the general
procedure, 62 mg of dihydro ferulic acid (0.315 mmol, 1.05
eq) were used. Column chromatography (CH2Cl2 : MeOH =
30:1) followed by preparative HPLC (Method C) gave 35
mg (0.05 mmol, 18 %) of the desired product as a white
solid. HPLC-purity: ≥ 99 % 1H-NMR (400 MHz, CDCl3)
δ(ppm) = 2.82 – 2.86 (m, 2 H, 2’-CH2), 2.97 (t, J = 7.3 Hz,
2 H, 3’-CH2), 3.51 (br. s., 1H, 3-OH), 3.56 (dd, J = 12.4,
3.9 Hz, 1 H, 23a-H), 3.80 (dd, J = 12.3, 2.8 Hz, 1 H, 23b-
H), 3.87 (s, 3 H, OMe), 3.91 (s, 3 H, OMe), 4.05 (dddd, J =
8.5, 3.8, 2.8, 1.3 Hz, 1 H, 10-H), 4.57 (dd, J = 12.0, 5.0 Hz,
1 H, 3-H), 4.95 (d, J = 8.3 Hz, 1 H, 11-H), 5.04 (d, J = 12.0
Hz, 1 H, 2-H), 6.24 (dd, J = 5.0, 2.0 Hz, 1 H, 8-H), 6.30
(dd, J = 2.0, 0.5 Hz, 1 H, 6-H), 5.57 (br. s., 1H, 20-OH),
5.78 (br. s., 1H, d-OH), 6.72 – 6.74 (m, 2 H, b-H, f-H), 6.84
– 6.87 (m, 1 H, e-H), 6.92 – 6.95 (m, 3 H, 18-H, 21-H, 22-
H), 7.03 – 7.07 (m, 2 H, 15-H, 16-H), 7.17 (dd, J = 7.0, 1.8
Hz, 1 H, 13-H), 11.06 (d, J = 1.8 Hz, 1 H, 5-OH). 13C-
NMR (100 MHz, CDCl3) δ(ppm) = 30.5 (3’-CH2), 36.4
(2’-CH2), 55.9 (OMe), 56.1 (OMe), 61.7 (23-CH2),
72.60/72.68 (3-CH), 76.37/76.41 (11-CH), 78.33/78.37 (10-
CH), 83.13/83.17 (2-CH), 102.1 (8-CH), 103.6 (6-CH),
104.1 (4a-Cq), 109.56/109.58 (18-CH), 111.0 (b-CH), 114.5
(e-CH), 114.70/114.72 (21-CH), 116.51/116.63 (13-CH),
117.27/117.37 (16-CH), 120.8 (22-CH), 120.9 (f-CH),
120.95/121.15 (15-CH), 127.8 (17-Cq), 129.00/129.04 (14-
Cq), 131.7 (a-Cq), 143.6 (12a-Cq), 144.24/144.27 (16-Cq),
144.3 (d-Cq), 146.5 (c-Cq), 146.6 (20-Cq), 147.0 (19-Cq),
159.2 (8a-Cq), 162.29/162.31 (7a-Cq), 162.8 (5-OH), 170.3
(1’-COO), 197.4 (4-CO). ESI-MS: 661.20 m/z [M+H]+,
683.10 m/z [M+Na]+.
7-O-Sinapinoylsilibinin (7a) Following the general
procedure, 70 mg of sinapic acid (0.315 mmol, 1.05 eq)
were used. Column chromatography (CH2Cl2 : MeOH =
75:1 ‰ 65:1 ‰ 50:1) followed by preparative HPLC
(Method D) gave 22 mg (0.03 mmol, 11 %) of the desired
product as a yellow solid. HPLC-purity: 98.6 % 1H-NMR
(400 MHz, CDCl3) δ(ppm) = 2.02 (br. s., 1H, 23-OH), 3.50
(br. s., 1 H, 3-OH), 3.56 (dd, J = 12.4, 2.9 Hz, 1 H, 23a-H),
3.81 (dd, J = 12.3, 2.3 Hz, 1 H, 23b-H), 3.91 (s, 3 H, OMe),
3.93 (s, 6 H, OMe), 4.05 (dddd, J = 8.3, 4.0, 2.8, 1.3 Hz, 1
H, 10-H), 4.60 (dd, J = 12.0, 5.3 Hz, 1 H, 3-H), 4.95 (d, J =
8.3 Hz, 1 H, 11-H), 5.06 (d, J = 12.3 Hz, 1 H, 2-H), 5.76
(br. s., 1 H, 20-OH), 5.88 (br. s., 1 H, d-OH), 6.39 (dd, J =
5.5, 2.0 Hz, 1 H, 8-H), 6.43 (dd, J = 15.3, 1.3 Hz, 1H, 2’-
H), 6.45 (d, J = 2.0 Hz, 1 H, 6-H), 6.82 (s, 2 H, b-H, f-H),
6.92 – 6.96 (m, 3 H, 18-H, 21-H, 22-H), 7.03 – 7.10 (m, 2
H, 15-H, 16-H), 7.19 (dd, J = 7.4, 1.9 Hz, 1 H, 13-H), 7.77
(d, J = 15.6 Hz, 1 H, 3’-H), 11.09 (br. s., 1 H, 5-OH). 13C-
NMR (100 MHz, CDCl3) δ(ppm) = 56.1 (OMe), 56.4
(2xOMe), 61.7 (23-CH2), 72.61/72.70 (3-CH), 76.36/76.41
(11-CH), 78.33/78.37 (10-H), 83.14/83.17 (2-CH), 102.2
(8-CH), 103.6 (6-CH), 104.0 (4a-Cq), 105.5 (b-CH, f-CH),
109.53/109.56 (18-CH), 114.0 (2’-CH), 114.7 (21-CH),
116.51/116.64 (13-CH), 117.27/117.38 (16-CH), 120.8 (22-
CH), 120.94/121.17 (15-CH), 125.4 (a-Cq), 127.8 (17-Cq),
129.06/129.10 (14-Cq), 137.9 (d-Cq), 143.95/143.96 (12a-
Cq), 144.2 (16a-Cq), 146.5 (20-Cq), 147.0 (c-Cq, e-Cq),
147.3 (19-Cq), 148.1 (3’-CH), 159.5 (8a-Cq), 162.32/162.34
(7-Cq), 162.79/162.82 (5-OH), 164.3 (1’-COO), 197.3 (4-
CO). ESI-MS: 689.25 m/z [M+H]+, 711.25 m/z [M+Na]+.
7-O-Vanillylsilibinin (6c) Following the general procedure,
63 mg of vanillic acid (0.315 mmol, 1.05 eq) were used.
Column chromatography (CH2Cl2 : MeOH = 45:1) followed
by preparative HPLC (Method C) gave 35 mg (0.06 mmol,
20 %) of the desired product as a white solid. HPLC-
purity: 96.5 % 1H-NMR (400 MHz, CDCl3) δ(ppm) =
3.56 (dd, J = 12.5, 4.0 Hz, 1 H, 23a-H), 3.81 (dd, J = 12.3,
2.8 Hz, 1 H, 23b-H), 3.92 (s, 3 H, OMe), 3.98 (s, 3 H,
OMe), 4.05 (dddd, J = 8.2, 3.9, 2.6, 1.6 Hz, 1 H, 10-H),
4.61 (dd, J = 12.0, 5.5 Hz, 1 H, 3-H), 4.96 (d, J = 8.3 Hz, 1
H, 11-H), 5.08 (d, J = 12.0 Hz, 1 H, 2-H), 5.75 (br. s., 1H,
20-OH), 6.19 (br. s., 1H, d-OH), 6.42 (dd, J = 5.0, 2.0 Hz, 1
H, 8-H), 6.48 (dd, J = 2.0, 0.5 Hz, 1 H, 6-H), 6.92 – 6.96
(m, 3 H, 18-H, 21-H, 22-H), 7.00 (dd, J = 8.3, 1.0 Hz, 1 H,
e-H), 7.05 – 7.10 (m, 2 H, 15-H, 16-H), 7.19 (dd, J = 7.2,
1.9 Hz, 1 H, 13-H), 7.61 (dd, J = 1.8, 1.0 Hz, 1 H, b-H),
7.77 (ddd, J = 8.3, 2.0, 1.0 Hz, 1 H, f-H), 11.09 (d, J = 1.5
Hz, 1 H, 5-OH). 13C-NMR (100 MHz, CDCl3) δ(ppm) =
56.1 (OMe), 56.3 (OMe), 61.7 (23-CH2), 72.64/72.72 (3-
CH), 76.37/76.41 (11-CH), 78.33/78.36 (10-CH),
83.15/83.18 (2-CH), 102.4 (8-CH), 103.9 (6-CH), 104.1
(4a-Cq), 109.52/109.55 (18-CH), 112.3 (b-CH), 114.4 (e-
CH), 114.68/114.69 (21-CH), 116.52/116.65 (13-CH),
117.28/117.38 (16-CH), 120.6 (a-Cq), 120.8 (22-CH),
120.93/121.15 (15-CH), 125.4 (f-CH), 127.8 (17-Cq),
7-O-Dihydrosinapinoylsilibinin (7b) Following the
general procedure, 71 mg of dihydrosinapic acid (0.315
mmol, 1.05 eq) were used. Column chromatography
(CH2Cl2 : MeOH = 75:1 ‰ 65:1 ‰ 50:1) gave 104 mg
(0.15 mmol, 50 %) of the desired product as a white solid.
HPLC-purity: 97.0 % 1H-NMR (400 MHz, CDCl3)
δ(ppm) = 2.81 – 2.87 (m, 2 H, 3’-H), 2.93 – 3.00 (m, 2 H,
2’-H), 3.51 – 3.56 (m, 1 H, 23a-H), 3.57 (br. s, 1 H, OH),
3.75 – 3.83 (m, 1 H, 23b-H), 3.86 (s, 3 H, OMe), 3.86 (s, 3
H, OMe), 3.89 (s, 3 H, OMe), 4.03 (dtd, J = 8.5, 2.8, 2.8,
1.0 Hz, 1 H, 10-H), 4.56 (ddd, J = 12.1, 5.0, 2.0 Hz, 1 H, 3-
H), 4.93 (d, J = 8.3 Hz, 1 H, 11-H), 5.02 (d, J = 12.3 Hz, 1
H, 2-H), 5.50 (br. s, 1 H, d-OH), 5.84 (s, 1 H, 20-OH), 6.24
(dd, J = 4.8, 2.0 Hz, 1 H, 6-H), 6.30 (dd, J = 2.0, 0.8 Hz, 1
H, 8-H), 6.45 (d, J = 0.8 Hz, 2 H, b-H, f-H), 6.90 – 6.95 (m,
3 H, 18-H, 21-H, 22-H), 7.00 – 7.08 (m, 2 H, 15-H, 16-H),
7.16 (dd, J = 6.4, 1.9 Hz, 1 H, 13-H), 11.08 (d, J = 2.5 Hz, 1
H, 5-OH). 13C-NMR (100 MHz, CDCl3) δ(ppm) = 31.0
(2’-CH2), 36.4 (3’-CH2), 56.07 (OMe), 56.33 (2xOMe),
61.7 (23-CH2), 72.60/72.67 (3-CH), 76.35/76.39 (11-CH),
78.35/78.38 (10-H), 83.12/83.16 (2-CH), 102.1 (8-CH),
103.5 (6-CH), 104.1 (4a-Cq), 105.1 (b-CH, f-CH),
109.60/109.64 (18-CH), 114.7 (21-CH), 116.52/116.62 (13-
CH),
117.25/117.33
(16-CH),
120.8
(22-CH),
120.93/121.12 (15-CH), 127.8 (17-Cq), 129.01/129.04 (14-
Cq), 130.9 (a-Cq), 133.4 (d-Cq), 143.92/143.94 (12a-Cq),
144.23/144.25 (16a-Cq), 146.4 (20-Cq), 147.0 (19-Cq),
147.1 (c-Cq, e-Cq), 159.1 (8a-Cq), 162.29/162.31 (7-Cq),
129.06/129.11
(14-Cq),
143.94/143.96
(12a-Cq),
144.22/144.26 (16a-Cq), 146.4 (20-Cq, c-Cq), 146.96/146.97
(19-Cq), 151.1 (d-Cq), 159.8 (8a-Cq), 162.4 (7a-Cq), 162.8

ACCEPTED MANUSCRIPT
162.76/162.78 (5-OH), 170.2 (1’-COO), 197.4 (4-CO).
147.9 (3’-CH), 148.3 (c-Cq), 150.3 (d-Cq), 159.5 (8a-Cq),
162.27/162.29 (7-Cq), 162.76/162.77 (5-OH), 164.4 (1’-
COO), 197.4 (4-CO). ESI-MS: 707.30 m/z [M+Na]+.
ESI-MS: 691.35 m/z [M+H]+, 713.25 m/z [M+Na]+.
7-O-Syringoylsilibinin (7c) Following the general
procedure, 62 mg of sinapic acid (0.315 mmol, 1.05 eq)
were used. Column chromatography (CH2Cl2 : MeOH =
75:1‰ 65:1 ‰ 50:1) followed by preparative HPLC
(Method D) gave 49 mg (0.07 mmol, 25 %) of the desired
7-O-[3,4-O-(1-Methylethylidene)]
dihydrocaffeoylsilibinin (8b) Following the general
procedure, 70 mg of 3,4-O-(1-methylethylidene) dihydro
caffeic acid (0.315 mmol, 1.05 eq) were used. Column
chromatography (CH2Cl2 : MeOH = 45:1) followed by
preparative HPLC (Method B) gave 81 mg (0.12 mmol, 40
%) of the desired product as a white solid. HPLC-purity:
1
product as a white solid. HPLC-purity: 98.8 % H-NMR
(400 MHz, CDCl3) δ(ppm) = 1.99 (br. s., 1H, 23-OH), 3.49
(s, 1 H, 3-OH), 3.57 (dd, J = 12.5, 3.8 Hz, 1 H, 23a-H), 3.81
(dd, J = 12.4, 2.6 Hz, 1 H, 23b-H), 3.92 (s, 3 H, OMe), 3.97
(s, 3 H, OMe), 3.97 (s, 3 H, OMe), 4.05 (dddd, J = 8.3, 4.0,
2.6, 1.6 Hz, 1 H, 10-H), 4.62 (dd, J = 12.2, 5.4 Hz, 1 H, 3-
H), 4.96 (d, J = 8.3 Hz, 1 H, 11-H), 5.08 (d, J = 12.3 Hz, 1
H, 2-H), 5.73 (br. s., 1 H, 20-OH), 6.07 (br. s., 1 H, d-OH),
6.43 (dd, J = 5.1, 2.1 Hz, 1 H, 8-H), 6.49 (dd, J = 2.0, 0.8
Hz, 1 H, 6-H), 6.93 – 6.96 (m, 3 H, 18-H, 21-H, 22-H), 7.05
– 7.10 (m, 2 H, 15-H, 16-H), 7.20 (dd, J = 7.5, 2.0 Hz, 1 H,
13-H), 7.41 (d, J = 1.0 Hz, 2 H, b-H, f-H), 11.10 (d, J = 1.3
Hz, 1 H, 5-OH). 13C-NMR (100 MHz, CDCl3) δ(ppm) =
56.1 (OMe), 56.6 (2xOMe), 61.7 (23-CH2), 72.65/72.73 (3-
1
95.6 % H-NMR (400 MHz, CDCl3) δ(ppm) = 1.66 (s, 6
H, 2xMe), 2.23 (br.s., 1 H, 23-OH), 2.81 (t, J = 7.8 Hz, 2 H,
2’-H), 2.94 (t, J = 7.3 Hz, 2 H, 3’-H), 3.54 (dd, J = 12.3, 3.5
Hz, 1 H, 23a-H), 3.62 (br. s., 1H, 3-OH), 3.79 (dd, J = 12.3,
1.8 Hz, 1 H, 23b-H), 3.89 (s, 3 H, OMe), 4.03 (dt, J = 7.8,
3.2 Hz, 1 H, 10-H), 4.55 (dd, J = 12.0, 4.5 Hz, 1 H, 3-H),
4.93 (d, J = 8.0 Hz, 1 H, 11-H), 5.02 (d, J = 12.0 Hz, 1 H,
2-H), 5.88 (br. s., 1H, 20-OH), 6.23 (dd, J = 4.4, 2.1 Hz, 1
H, 8-H), 6.30 (dd, J = 2.0, 0.8 Hz, 1 H, 6-H), 6.61 – 6.66
(m, 3 H, b-H, e-H, f-H), 6.90 – 6.96 (m, 3 H, 18-H, 21-H,
22-H), 7.00 – 7.07 (m, 2 H, 15-H, 16-H), 7.16 (dd, J = 5.8,
1.8 Hz, 1 H, 13-H), 11.08 (d, J = 2.5 Hz, 1 H, 5-OH). 13C-
NMR (100 MHz, CDCl3) δ(ppm) = 25.8 (2xMe), 30.6 (3’-
CH), 36.3 (2’-CH), 56.1 (OMe), 61.6 (23-CH2),
72.60/72.67 (3-CH), 76.34/76.37 (11-CH), 78.35/78.38 (10-
CH), 83.11/83.14 (2-CH), 102.1 (8-CH), 103.6 (6-CH),
104.1 (4a-Cq), 108.2 (b-CH), 108.7 (e-CH), 109.64/109.68
(18-CH), 114.8 (21-CH), 116.57/116.64 (13-CH),
117.23/117.32 (16-CH), 117.9 (C(CH3)2), 120.6 (22-CH),
120.8 (f-CH), 120.99/121.14 (15-CH), 127.8 (17-Cq),
129.05/129.08 (14-Cq), 132.8 (a-Cq), 143.9 (12a-Cq),
144.21/144.23 (16a-Cq), 146.1 (20-Cq), 146.4 (c-Cq), 147.0
(19-Cq), 147.6 (d-Cq), 159.1 (8a-Cq), 162.25/162.28 (7-Cq),
162.75/162.76 (5-OH), 170.2 (1’-COO), 197.5 (4-CO).
ESI-MS: 687.00 m/z [M+H]+, 709.05 m/z [M+Na]+.
CH), 76.37/76.41
(11-CH), 78.33/78.36
(10-H),
83.16/83.20 (2-CH), 102.4 (8-CH), 103.9 (6-CH), 104.1
(4a-Cq), 107.5 (b-CH, f-CH), 109.50/109.53 (18-CH),
114.67/114.69
(21-CH),
116.50/116.65
(13-CH),
117.28/117.38 (16-CH), 119.43 (a-Cq), 120.8 (22-CH),
120.91/121.15 (15-CH), 127.8 (17-Cq), 129.04/129.09 (14-
Cq), 140.4 (d-Cq), 144.24/144.27 (12a-Cq, 16a-Cq), 146.5
(20-Cq), 146.9 (c-Cq, e-Cq), 147.0 (19-Cq), 159.7 (8a-Cq),
162.4 (7-Cq), 162.9 (5-OH), 163.7 (a-COO), 197.38/197.40
(4-CO). ESI-MS: 663.25 m/z [M+H]+, 685.30 m/z
[M+Na]+.
7-O-[3,4-O-(1-Methylethylidene)]caffeoylsilibinin
(8a)
Following the general procedure 69 mg 3,4-O-(1-
methylethylidene)caffeic acid (0.315 mmol, 1.05 eq) were
used. Column chromatography (CH2Cl2 : MeOH = 50:1)
gave 130 mg (0.19 mmol, 63 %) of the desired product as a
7-O-[3,4-O-(1-Methylethylidene)] 3,4-dihydroxybenzyl-
silibinin (8c) Following the general procedure 61 mg of
3,4-O-(1-Methylethylidene) 3,4-dihydroxybenzoic acid
(0.315 mmol, 1.05 eq) were used. Column chromatography
(CH2Cl2 : MeOH = 50:1) followed by preparative HPLC
gave 73 mg (0.11 mmol, 37 %) of the desired product as a
1
white solid. HPLC-purity: 95.6 % H-NMR (400 MHz,
CDCl3) δ (ppm) = 1.70 (s, 6 H, 2xMe), 3.54 (dd, J = 12.5,
4.0 Hz, 1 H, 23a-H), 3.79 (dd, J = 12.5, 2.5 Hz, 1 H, 23b-
H), 3.89 (s, 3 H, OMe), 4.02 (ddd, J= 8.3, 3.8, 2.8 Hz, 1 H,
10-H), 4.56 (dd, J = 12.0, 5.5 Hz, 1 H, 3-H), 4.92 (dd, J =
8.3, 1.8 Hz, 1 H, 11-H), 5.04 (d, J = 12.0 Hz, 1 H, 2-H),
6.34 (dd, J = 16.1, 1.3 Hz, 1 H, 2’-H), 6.38 (dd, J = 4.6, 2.1
Hz, 1 H, 8-H), 6.44 (d, J = 2.0 Hz, 1 H, 6-H), 6.74 (d, J =
7.8 Hz, 1 H, b-H), 6.91 – 6.94 (s, 3 H, 18-H, 21-H, 22-H),
6.98 – 7.03 (m, 3 H, 16-H, e-H, f-H), 7.04 – 7.08 (m, 1 H,
15-H), 7.17 (dd, J = 5.8, 2.0 Hz, 1 H, 13-H), 7.75 (dd, J =
15.8, 0.5 Hz, 1 H, 3’-H), 11.10 (d, J = 2.8 Hz, 1 H, 5-OH).
13C-NMR (100 MHz, CDCl3) δ (ppm) = 25.9 (2xMe), 56.1
(OMe), 61.6 (23-CH2), 72.62/72.68 (3-CH), 76.33/76.36
(11-CH), 78.35/78.39 (10-CH), 83.11/83.14 (2-CH), 102.2
(8-CH), 103.7 (6-CH), 104.0 (4a-Cq), 106.5 (e-CH), 108.5
(b-CH), 109.63/109.67 (18-CH), 113.5 (2’-CH),
1
white solid. HPLC-purity: ≥ 99 % H-NMR (400 MHz,
CDCl3) δ (ppm) = 1.72 (s, 6 H, 2xMe), 3.55 (dd, J = 12.5,
4.0 Hz, 1 H, 23a-H), 3.79 (dd, J = 12.5, 2.8 Hz, 1 H, 23b-
H), 3.90 (s, 3 H, OMe), 4.03 (ddd, J = 8.3, 3.8, 2.8 Hz, 1 H,
10-H), 4.58 (dd, J = 12.2, 5.6 Hz, 1 H, 3-H), 4.93 (dd, J =
8.3, 1.5 Hz, 1 H, 11-H), 5.05 (d, J = 12.0 Hz, 1 H, 2-H),
5.83 (br. s., 1H, 20-OH), 6.40 (dd, J = 5.0, 2.3 Hz, 1 H, 8-
H), 6.47 (d, J = 1.8 Hz, 1 H, 6-H), 6.80 (dd, J = 8.3, 1.0 Hz,
1 H, f-H), 6.92 – 6.94 (m, 3 H, 18-H, 21-H, 22-H), 7.01 –
7.07 (m, 2 H, 15-H, 16-H), 7.18 (dd, J = 5.8, 2.0 Hz, 1 H,
13-H), 7.47 (dd, J = 1.6, 1.1 Hz, 1 H, b-H), 7.73 (dq, J =
8.3, 1.0 Hz, 1 H, e-H), 11.11 (d, J = 2.5 Hz, 1 H, 4-OH).
13C-NMR (100 MHz, CDCl3) δ (ppm) = 25.9 (2xMe), 56.1
(OMe), 61.7 (23-CH2), 72.63/72.70 (3-CH), 76.33/76.38
(11-CH), 78.35/78.38 (10-CH), 83.13/83.17 (2-CH), 102.3
(8-CH), 103.8 (6-CH), 104.1 (4a-Cq), 108.1 (f-CH),
109.59/109.63 (18-CH), 109.8 (b-CH), 114.7 (21-CH),
116.56/116.64 (13-CH), 117.25/117.33 (16-CH), 119.8
114.73/114.75
(21-CH),
116.57/116.65
(13-CH),
117.24/117.32 (16-CH), 119.3 (C(CH3)2), 120.8 (22-CH),
120.99/121.15 (15-CH), 125.0 (f-CH), 127.8 (a-Cq), 127.9
(17-Cq), 129.12/129.15 (14-Cq), 143.9 (12a-Cq),
144.19/144.22 (16a-Cq), 146.4 (20-Cq), 147.0 (19-Cq),

ACCEPTED MANUSCRIPT
(C(CH3)2), 120.8 (22-CH), 120.98/121.15 (15-CH), 121.8
(a-Cq), 126.2 (e-CH), 127.8 (17-Cq), 129.09/129.13 (14-Cq),
143.9 (12a-Cq), 144.2 (16a-Cq), 146.42/146.43 (20-Cq),
147.0 (19-Cq), 147.8 (c-Cq), 152.5 (d-Cq), 159.7 (8a-Cq),
162.31/162.33 (7-Cq), 162.79/162.81 (5-OH), 163.5 (a-
COO), 197.4 (4-CO). ESI-MS: 659.25 m/z [M+H]+, 681.15
m/z [M+Na]+.
117.30/117.40 (16-CH), 119.6 (a-Cq), 120.8 (22-CH),
120.93/121.16 (15-CH), 126.6 (a’-Cq), 127.8 (17-Cq), 129.4
(14-Cq), 134.0 (d’-Cq), 140.0 (d-Cq), 144.0 (12a-Cq),
144.27/144.30 (16a-Cq), 146.5 (20-Cq), 146.8 (c’-Cq, e’-Cq),
147.0 (19-Cq), 152.7 (c-Cq, e-Cq), 159.4 (8a-Cq),
162.41/162.44 (7-Cq), 162.88/162.90 (5-OH), 163.4 (a-
COO), 163.8 (a’-COO), 197.5 (4-CO). ESI-MS: 843.35
m/z [M+H]+.
23-O-Cinnamoylsilibinin (9) Following the general
procedure, 47 mg of cinnamic acid (0.315 mmol, 1.05 eq)
and 47 µL triethylamine (0.33 mmol, 1.1 eq) were used.
Column chromatography (CH2Cl2 : MeOH = 50:1) gave 70
mg (0.11 mmol, 37 %) of the product as a white solid.
HPLC-purity: 98.9 % 1H-NMR (400 MHz, CDCl3)
δ(ppm) = 3.87 (s, 3 H, OMe), 4.13 (ddd, J = 12.1, 4.5, 0.8
Hz, 1 H, 23a-H), 4.28 (ddt, J = 7.9, 4.7, 2.9, 2.9 Hz, 1 H,
10-H), 4.39 (dd, J = 12.3, 2.8 Hz, 1 H, 23b-H), 4.49 (dd, J =
11.8, 3.5 Hz, 1 H, 3-H), 4.91 (dd, J = 8.0, 1.5 Hz, 1 H, 11-
H), 4.96 (d, J = 12.0 Hz, 1 H, 2-H), 5.94 (dd, J = 3.6, 2.1
Hz, 1 H, 8-H), 6.02 (dd, J = 2.1, 1.1 Hz, 1 H, 6-H), 6.44
(dd, J = 16.1, 1.5 Hz, 1 H, 2’-H), 6.85 – 6.95 (m, 3 H, 18-H,
21-H, 22-H), 7.02 – 7.09 (m, 2 H, 15-H, 16-H), 7.17 (dd, J
= 4.3, 1.8 Hz, 1 H, 13-H), 7.35 – 7.44 (m, 3 H, b-H, d-H, f-
H), 7.50 – 7.55 (m, 2 H, c-H, e-H), 7.66 (d, J = 16.1 Hz, 1
H, 3’-H), 11.22 (br. s, 1 H, 5-OH). 13C-NMR (100 MHz,
CDCl3) δ(ppm) = 56.0 (OMe), 63.1 (23-CH2), 72.27/72.34
(3-CH), 75.94/75.97 (10-CH), 76.56/76.63 (11-CH),
82.92/82.95 (2-CH), 96.0 (8-CH), 97.1 (6-CH), 100.6 (4a-
Cq), 109.44/109.46 (18-CH), 114.9 (21-CH), 116.53/116.59
(13-CH), 117.0 (2’-CH), 117.44/117.52 (16-CH), 120.8
(22-CH), 121.18/121.31 (15-CH), 127.5 (17-Cq), 128.3 (c-
CH, e-CH), 129.0 (b-CH, f-CH), 129.49/129.50 (14-Cq),
130.7 (d-CH), 134.1 (a-Cq), 143.76/143.77 (12a-Cq), 144.0
(16a-Cq), 145.9 (3’-CH), 146.6 (20-Cq), 147.1 (19-Cq),
163.1 (8a-Cq), 163.8 (5-OH), 166.5 (7-OH), 166.7 (1’-
COO), 195.6 (4-CO). ESI-MS: 613.10 m/z [M+H]+, 635.05
m/z [M+Na]+.
7-O-Feruloylsilybin
A (11a) Following the general
procedure, 46 mg of ferulic acid (0.22 mmol, 1.05 eq) were
used. Column chromatography (CH2Cl2 : MeOH = 75:1 ‰
65:1 ‰ 50:1) followed by preparative HPLC (Method E)
gave 30 mg (0.05 mmol, 23 %) of the desired product as a
1
yellow solid. HPLC-purity: 99.2 % H-NMR (400 MHz,
CDCl3) δ(ppm) = 3.46 (br. s., 1 H, 3-OH), 3.57 (br. d, J =
12.3 Hz, 1 H, 23a-H), 3.81 (dd, J = 12.4, 2.4 Hz, 1 H, 23b-
H), 3.93 (s, 1 H, OMe), 3.95 (s, 3 H, OMe), 4.06 (ddd, J =
8.3, 4.0, 2.8 Hz, 1 H, 10-H), 4.62 (d, J = 12.3 Hz, 1 H, 3-
H), 4.97 (d, J = 8.3 Hz, 1 H, 11-H), 5.07 (d, J = 12.0 Hz, 1
H, 2-H), 5.72 (br. s., 1 H, 20-OH), 5.94 (br. s., 1 H, d-OH),
6.39 (d, J = 2.0 Hz, 1 H, 8-H), 6.41 (d, J = 16.1 Hz, 1 H, 2’-
H), 6.45 (d, J = 2.0 Hz, 1 H, 6-H), 6.92 – 6.97 (m, 4 H, 18-
H, 21-H, 22-H, e-H), 7.04 – 7.11 (m, 3 H, 15-H, 16-H, b-
H), 7.14 (dd, J = 8.5, 2.0 Hz, 1 H, f-H), 7.21 (d, J = 1.8 Hz,
1 H, 13-H), 7.79 (d, J = 15.8 Hz, 1 H, 3’-H), 11.07 (s, 1 H,
5-OH). 13C-NMR (100 MHz, CDCl3) δ(ppm) = 56.02
(OMe), 56.07 (OMe), 61.7 (23-CH2), 72.6 (3-CH), 76.4
(11-CH), 78.3 (10-CH), 83.2 (2-CH), 102.2 (8-CH), 103.7
(6-CH), 104.7 (4a-Cq), 109.5 (18-CH), 109. 6 (b-CH), 113.6
(2’-CH), 114.7 (21-CH), 114.9 (e-CH), 116.5 (13-CH),
117.3 (16-CH), 120.8 (22-CH), 121.2 (15-CH), 123.8 (f-
CH), 126.5 (a-Cq), 127.8 (17-Cq), 129.1 (14-Cq), 144.0
(12a-Cq), 144.3 (16a-Cq), 146.5 (20-Cq), 146.9 (19-Cq),
147.0 (c-Cq), 147.9 (3’-CH), 148.7 (d-Cq), 159.6 (7-Cq),
162.4 (8a-Cq), 162.8 (5-Cq), 164.3 (1’-COO), 197.3 (4-CO).
ESI-MS: 659.25 m/z [M+H]+.
7-O-Feruloylsilybin
B (11b) Following the general
7-O-[4-((4-hydroxy-3,5-dimethoxybenzoyl)oxy)-3,5-
dimethoxybenzoyl]silibinin (10) Byproduct in the
preparation of compound 7c. Column chromatography
(CH2Cl2 : MeOH = 75:1 ‰ 65:1 ‰ 50:1) followed by
preparative HPLC (Method D) gave 17 mg (0.02 mmol, 7
%) of the diester as a white solid. HPLC-purity: 97.3 %
1H-NMR (400 MHz, CDCl3) δ(ppm) = 2.00 (br. s., 1 H,
23-OH), 3.50 (br. s., 1 H, 3-OH), 3.57 (dd, J = 12.5, 3.8 Hz,
1 H, 23a-H), 3.81 (dd, J = 12.4, 2.6 Hz, 1 H, 23b-H), 3.89
(s, 3 H, OMe), 3.90 (m, 3 H, OMe), 3.92 (s, 3 H, OMe),
3.97 (s, 6 H, 2xOMe), 4.06 (dddd, J = 8.2, 3.9, 2.6, 1.1 Hz,
1 H, 10-H), 4.63 (dd, J = 12.0, 5.3 Hz, 1 H, 3-H), 4.96 (d, J
= 8.3 Hz, 1 H, 11-H), 5.09 (d, J = 12.0 Hz, 1 H, 2-H), 5.75
(br. s, 1 H, 20-OH), 6.04 (br. s, 1 H, d’-OH), 6.44 (dd, J =
4.8, 2.0 Hz, 1 H, 8-H), 6.51 (dd, J = 2.0, 0.8 Hz, 1 H, 6-H),
6.93 – 6.96 (m, 3 H, 18-H, 21-H, 22-H), 7.04 – 7.10 (m, 2
H, 15-H, 16-H), 7.20 (dd, J = 7.2, 1.9 Hz, 1 H, 13-H), 7.46
(d, J = 1.0 Hz, 2 H, b-H, f-H), 7.51 (s, 2 H, b’-H, f’-H),
11.12 (d, J = 1.8 Hz, 1 H, 5-OH). 13C-NMR (100 MHz,
CDCl3) δ(ppm) = 56.1 (OMe), 56.49 (2xOMe), 56.53
(2xOMe), 61.7 (23-CH2), 72.65/72.74 (3-CH), 76.38/76.42
(11-CH), 78.35/78.38 (10-H), 83.20/83.24 (2-CH), 102.4
(8-CH), 103.9 (6-CH), 104.2 (4a-Cq), 107.1 (b’-CH, f’-CH),
procedure, 46 mg of ferulic acid (0.22 mmol, 1.05 eq) were
used. Column chromatography (CH2Cl2 : MeOH = 75:1 ‰
65:1 ‰ 50:1) followed by preparative HPLC (Method E)
gave 36 mg (0.06 mmol, 27 %) of the desired product as a
1
yellow solid. HPLC-purity: 97.2 % H-NMR (400 MHz,
CDCl3) δ(ppm) = 1.97 (br. s., 1 H, 23-OH), 3.48 (br. s., 1
H, 3-OH), 3.57 (br. d, J = 12.3 Hz, 1 H, 23a-H), 3.81 (br.
dd, J = 12.4, 2.4 Hz, 1 H, 23b-H), 3.92 (s, 3 H, OMe), 3.95
(s, 3 H, OMe), 4.05 (ddd, J = 8.3, 4.0, 2.8 Hz, 1 H, 10-H),
4.60 (d, J = 12.0 Hz, 1 H, 3-H), 4.97 (d, J = 8.3 Hz, 1 H,
11-H), 5.07 (d, J = 12.0 Hz, 1 H, 2-H), 5.73 (br. s., 1 H, 20-
OH), 5.96 (br. s., 1 H, d-OH), 6.40 (d, J = 2.0 Hz, 1 H, 8-
H), 6.41 (d, J = 15.8 Hz, 1 H, 2’-H), 6.45 (d, J = 2.0 Hz, 1
H, 6-H), 6.93 – 6.97 (m, 4 H, 18-H, 21-H, 22-H, e-H), 7.05
– 7.12 (m, 3 H, 15-H, 16-H, b-H), 7.14 (dd, J = 8.5, 2.0 Hz,
1 H, f-H), 7.19 (d, J = 2.0 Hz, 1 H, 13-H), 7.79 (d, J = 15.8
Hz, 1 H, 3’-H), 11.08 (s, 1 H, 5-OH). 13C-NMR (100 MHz,
CDCl3) δ(ppm) = 56.02 (OMe), 56.08 (OMe), 61.7 (23-
CH2), 72.7 (3-CH), 76.4 (11-CH), 78.4 (10-CH), 83.1 (2-
CH), 102.2 (8-CH), 103.7 (6-CH), 104.0 (4a-Cq), 109.5 (18-
CH), 109. 6 (b-CH), 113.6 (2’-CH), 114.7 (21-CH), 114.9
(e-CH), 116.7 (13-CH), 117.4 (16-CH), 120.8 (22-CH),
121.0 (15-CH), 123.8 (f-CH), 126.5 (a-Cq), 127.8 (17-Cq),
129.1 (14-Cq), 144.0 (12a-Cq), 144.2 (16a-Cq), 146.5 (20-
107.6
(b-CH,
f-CH),
109.51/109.54
116.52/116.65
(18-CH),
(13-CH),
114.68/114.70
(21-CH),

ACCEPTED MANUSCRIPT
Cq), 146.9 (19-Cq), 147.0 (c-Cq), 147.9 (3’-CH), 148.7 (d-
For neuroprotection 5 mM glutamate (monosodium-L-
glutamate, Sigma Aldrich, Munich, Germany) were
coincubated with 1, 5, 10 or 25 µM of respective
compounds for 24 hours. 25 µM quercetin (Sigma Aldrich,
Munich, Germany) served as positive control. After 24
hours incubation, cell viability was determined using the 3-
Cq), 159.6 (7-Cq), 162.3 (8a-Cq), 162.8 (5-Cq), 164.3 (1’-
COO), 197.3 (4-CO). ESI-MS: 659.20 m/z [M+H]+.
For detailed description of the chemical synthesis of all
compounds described in the article and spectral data refer to
the supporting information.
(4,5-dimethylthiazol-2-yl)-2,5-
diphenyl
tetrazolium
bromide (MTT, Sigma Aldrich, Munich, Germany) assay.
MTT solution (4 mg/mL in PBS) was diluted 1:10 with
medium and added to the wells after removal of previous
medium. Cells were incubated for 3 hours when the
supernatant was removed and lysis buffer (10% SDS) was
applied. The next day, absorbance at 560 nm was
determined with a multiwell plate photometer (Tecan –
SpectraMax 250). Experiments were performed in
triplicates.
4.2 FRAP Assay
The FRAP reagent was prepared by mixing 300 mmol/L
sodium acetate buffer (pH 3.6), 10 mmol/lL TPTZ (2,4,6-
tripyridyl-s-triazine, Sigma Aldrich, Munich, Germany)
solution in 40 mmol/L HCl and 20 mmol/L FeCl3·6H2O
solution at a 10:1:1 ratio. The freshly prepared FRAP
reagent and the SpectraMax 250 plate reader were warmed
to 37°C.
Results are presented as percentage to untreated control
cells. Data are expressed as means ± SD of three different
independent experiments. Analysis was accomplished using
GraphPad Prism 5 Software applying oneway ANOVA
followed by Dunnett’s multiple comparison posttest. Levels
of significance: * p < 0.05; ** p < 0.01; *** p < 0.001.
Levels of significance for Bonferroni’s posttest are
indicated as $.
0.1, 0.2, 0.4, 0.6, 0.8 and 1 mM methanolic solutions (10%
DMSO) of the compounds were prepared. FeSO4 was used
for calibration. Ascorbic acid, trolox (Sigma Aldrich,
Munich, Germany) and silibinin (Sigma Aldrich, Munich,
Germany) served as reference compounds.
In every well of a 96-well-plate 220.6 µL of FRAP reagent,
22 µL of distilled water, and 7.4 µL of sample or standard
were mixed and the absorbance changes were determined at
593 nm versus the FRAP reagent with methanol and DMSO
only as a blank. The final dilution of the test samples in the
reaction mixture was 1/34. The immediately starting
reaction was monitored every 15 seconds for 4 min with a
multiwell plate photometer (Tecan – SpectraMax 250).
Experiments were performed in triplicates.
4.4 Stability measurements
7-O-Feruloylsilibinin (6a) was dissolved in dimethyl
sulfoxide (DMSO) and diluted with either Dulbecco’s
Modified Eagle Medium (DMEM, Sigma Aldrich, Munich
Germany) or human pooled plasma (n = 6) to reach a final
concentration of 50 µM (1 % DMSO). The solution was
placed at 37°C with 5 % CO2 in an incubator (Thermo
Scientific) while shaking with 100 rpm (Unimax 1010
Schüttler, Heidolph, Schwalbach, Germany). Samples of
0.5 mL were taken after t = 0, 15, 30, 60, 120, 240, 480 min
and 24 h. The samples were treated with 0.5 mL of ice cold
acetonitrile. Precipitated proteins were removed by
centrifugation (Microfuge 22R, Beckman Coulter GmbH,
Krefeld, Germany) with 8000 g for 10 min at 4°C. The
solution was directly analysed by LCMS. (system from
Analysis was accomplished using GraphPad Prism 5
Software applying oneway ANOVA followed by
*
subsequent Tukey test. Significance was defined as p <
0.05.
4.3 Neuroprotection and neurotoxicity assay
HT-22 cells were grown in Dulbecco’s Modified Eagle
Medium (DMEM, Sigma Aldrich, Munich Germany)
supplemented with 10% (v/v) heat-inactivated fetal calf
serum (FCS) and 1% (v/v) penicillin-streptomycin. Cells
were subcultured every two days and incubated at 37°C
with 5% CO2 in a humidified incubator. Compounds were
dissolved in dimethyl sulfoxide (DMSO, Sigma Aldrich,
Munich, Germany) and diluted with medium.
Shimadzu Products equipped with
a
DGU-20A3R
controller, LC20AB liquid chromatograph, a SPD-20A
UV/Vis detector and an LCMS. Stationary phase was a
Synergi 4U fusion-RP (150 × 4.6 mm) column. Mobile
phase: gradient MeOH (0.01%TFA)/water (0.01%TFA)
(phase A/ phase B) were used. Gradient mode: 0 – 8 min
(10-80% phase A), 8 – 14 min (80% phase A), 14 – 14.3
min (80-90% phase A), 14.3 – 17 min (90% phase A), 17
min – 17.3 min (90-10% phase A) 17.3 – 18 min (10%
phase A).100 µL were injected and a flow rate of 1 mL/min
was used. ESI mass spectral data were acquired with a
Shimadzu LCMS-2020.) Signal areas were compared to the
signal area at t = 0 min, which was set as 100 %. The
experiment was carried out in triplicate.
Generally, 80% confluent cells were seeded with 5000 cells
per well into sterile 96-well plates and were incubated for
24 hours.
For the neurotoxicity assay, previous medium was removed
and 1, 5, 10 or 25 µM of the compound was added to the
wells. 0.05% DMSO in DMEM served as control. Cells
were incubated for 24 hours.

ACCEPTED MANUSCRIPT
[13] L. Fang, B. Kraus, J. Lehmann, J. Heilmann, Y.
Acknowledgment
We gratefully
diastereomerically pure silybin A and B by Professor
Vladimir Křen and Dr. David Biedermann (Centre of
Biotransformation
Microbiology. Czech Academy of Sciences, Prague, Czech
Republic).
Zhang, M. Decker, Bioorg. Med. Chem. Lett. 18 (2008)
2905-2909.
acknowledge
the
supply
of
[14] B. Kling, D. Bücherl, P. Palatzky, F.-M. Matysik, M.
Decker, J. Wegener, J. Heilmann, J. Nat. Prod. 77
(2014) 446-454.
and
Biocatalysis,
Institute
of
[15] E. Vavříková, V. Křen, L. Jezova-Kalachova, M.
Biler, B. Chantemargue, M. Pyszková, S. Riva, M.
Kuzma, K. Valentová, J. Ulrichová, J. Vrba, P.
Trouillas, J. Vacek, Eur. J. Med. Chem. 127 (2017)
263-274.
Supplementary data
Supplementary data related to this article can be found at …
…
[16] H.S. Althagafy, T.N. Graf, A.A. Sy-Cordero, B.T.
Gufford, M.F. Paine, J. Wagoner, S.J. Polyak, M.P.
Croatt, N.H. Oberlies, Bioorg. Med. Chem. 21 (2013)
3919-3926.
References
[17] D. Biedermann, E. Vavříková, L. Cvak and V. Křen,
[1] A.M. Palmer, Trends Pharmacol. Sci. 32 (2011) 141-
147.
Nat. Prod. Rep. 31 (2014) 1138-1175.
[18] a) X. Chen, K. Zenger, A. Lupp, B. Kling, J.
Heilmann, C. Fleck, B. Kraus, M. Decker, J. Med.
Chem. 55 (2012) 5231-5242.
[2] L. Ismaili, A. Romero, M. do Carmo Carreiras, J.
Marco-Contelles, Multitarget-directed antioxidants as
therapeutic agents: putting the focus on the oxidative
stress, in: M. Decker (Ed.), Design of hybrid molecules
for drug development, Elsevier, Oxford, UK (2017) 5-
46.
b) K. Zenger, X. Chen, M. Decker, B. Kraus, J.
Pharm. Pharmacol. 65 (2013) 1765-1772.
[19] R. Gažák, K. Valentová, K. Fuksová, P. Marhol, M.
Kuzma, M. Medina, I. Oborná, J. Ulrichová, V. Křen,
J. Med. Chem. 54 (2011) 7397-7407.
[3] L.X. Yang, K.X. Huang, H.B. Li, J.X. Gong, F. Wang,
Y.B. Feng, Q.F. Tao, Y.H. Wu, X.K. Li, X.M. Wu, S.
Zeng, S. Spencer, Y. Zhao, J. Qu, J. Med. Chem. 52
(2009) 7732-7752.
[20] R. Gažák, K. Purchtartová, P. Marhol, L. Živná, P.
Sedmera, K. Valentová, N. Kato, H. Matsumura, K.
Kaihatsu, V. Křen, Eur. J. Med. Chem. 45 (2010),
1059-1067.
[4] A. Zarrelli, A. Sgambato, V. Petito, L. De Napoli, L.
Previtera, G. Di Fabio, Bioorg. Med. Chem. Lett. 21
(2011) 4389-4392.
[21] P. Marhol, P. Bednář, P. Kolářová, R. Večeřa, J.
Ulrichová, E. Tesařová, E. Vavříková, M. Kuzma, V.
Křen, J. Funct. Foods 14 (2015) 570-580.
[5] F. Wang, K. Huang, L. Yang, J. Gong, Q. Tao, H. Li,
Y. Zhao, S. Zeng, X. Wu, J. Stöckigt, X. Li, J. Qu,
Bioorg. Med. Chem. 17 (2009) 6380-6389.
[22] M.F.M. Sciacca, V. Romanucci, A. Zarrelli, I.
Monaco, F. Lolicato, N. Spinella, C. Galati, G. Grasso,
L. D’Urso, M. Romeo, L. Diomede, M. Salmona, C.
Bongiorno, G. Di Fabio, C. La Rosa, D. Milardi, ACS
Chem. Neurosci. 8 (2017) 1767-1778.
[6] C. Pascual, R. Gonz, J. Armesto, P. Muriel, Drug Dev.
Res. 29 (1993) 73-77.
[7] a) R. Gažák, P. Sedmera, M. Marzorati, S. Riva, V.
Křen, J. Mol. Catal. B: Enzym. 50 (2008) 87-92.
[23] a) P.R. Davis-Searles, Y. Nakanishi, N. Kim, T.N.
Graf, N.H. Oberlies, M.C. Wani, M.E. Wall, R.
Agarwal, D.J. Kroll, Cancer Res. 65 (2005) 4448-4457.
b) P. Trouillas, P. Marsal, A. Svobodová, J. Vostálová,
R. Gažák, J. Hrbáč, P. Sedmera, V. Křen, R. Lazzaroni,
J. Duroux, D. Walterová, J. Phys. Chem. A 112 (2008)
1054-1063.
b) G. Deep, N.H. Oberlies, D.J. Kroll, R. Agarwal, Int.
J. Cancer 123 (2008) 41-50.
[8] R. Gažák, P. Sedmera, M. Vrbacký, J. Vostálová, Z.
Drahota, P. Marhol, D. Walterová, V. Křen, Free
Radical Biol. Med. 46 (2009) 745-758.
[24] M. Plíšková, J. Vondráček, V. Křen, R. Gažák, P.
Sedmera, D. Walterová, J. Psotová, V. Šimánek, M.
Machala, Toxicology 215 (2005) 80-89.
[9] I. György, S. Antus, G. Földiák, Radiat. Phys. Chem.
39 (1992) 81-84.
[25] K. Filippopoulou, N. Papaevgeniou, M. Lefaki, A.
Paraskevopoulou, D. Biedermann, V. Křen, N.
Chondrogianni, Free Radical Biol. Med. 103 (2017)
256-267.
[10] J. Heilmann, I. Calis, H. Kirmizibekmez, W. Schuhly,
S. Harput, O. Sticher, Planta Med. 66 (2000) 746-748.
[11] K. Ono, M. Hirohata, M. Yamada, Biochem. Biophys.
Res. Commun. 336 (2005) 444-449.
[26] K. Ishige, D. Schubert, Y. Sagara, Free Radic. Biol.
Med. 30 (2001) 433-446.
[12] J.J. Yan, J.Y. Cho, H.S. Kim, K.L. Kim, J.S. Jung,
S.O. Huh, H.W. Suh, Y.H. Kim, D.K. Song, Br. J.
Pharmacol. 133 (2001) 89-96.

ACCEPTED MANUSCRIPT
[27] a) A. Currais, C. Chiruta, M. Goujon-Svrzic, G. Costa,
[48] C. Ballard, S. Gauthier, A. Corbett, C. Brayne, D.
Aarsland, E. Jones, Lancet 377 (2011) 1019-1031.
T. Santos, M.T. Batista, J. Paiva, M. do Céu Madureira,
P. Maher, J. Ethnopharmacol. 155 (2014) 830-840.
[49] G. Invernizzi, E. Papaleo, R. Sabate, S. Ventura, Int. J.
Biochem. Cell Biol. 44 (2012) 1541-1554.
b) P. Maher, G.J. Kontoghiorghes, Neurochem. Res.
40 (2015) 609-620.
[50] R. Gažák, P. Trouillas, D. Biedermann, K. Fuksová, P.
Marhol, M. Kuzma, V. Křen, Tetrahedron Lett. 54
(2013) 315-317.
[28] A. Espargaró, A. Medina, O. Di Pietro, D. Muñoz-
Torrero, R. Sabate Sci. Rep. 6 (2016) 23349.
[29] A. Esparagó, T. Ginex, M. del Mar Vadell, M.A.
Busquets, J. Estelrich, D. Muñoz-Torrero, F.J. Luque,
R. Sabate J. Nat. Prod. 80 (2017) 278-289.
[51] S. Gunesch, S. Schramm, M. Decker, Future Med.
Chem. 9 (2017) 711-713.
[30] H. Koch, J. Tscherny, Arch. Pharm. 316 (1983) 426-
430.
[31] D. Monti, R. Gažák, P. Marhol, D. Biedermann, K.
Purchartová, M. Fedrigo, S. Riva, V. Křen, J. Nat.
Prod. 73 (2010) 613-619.
[32] P. Li, P.S. Callery, L. Gan, S.K. Balani Drug Metab.
Dispos. 35 (2007) 1203-1208.
[33] H. Sharp, J. Hollinshead, B.B. Bartholomew, J. Oben,
A. Watson, R.J. Nash, Nat. Prod. Commun. 2 (2007),
817-822.
[34] M. Novotná, R. Gažák, D. Biedermann, F. Di Meo, P.
Marhol, M. Kuzma, L. Bednárová, K. Fuksová, P.
Trouillas, V. Křen, Beilstein J. Org. Chem. 10 (2014)
1047-1063.
[35] W. Liu, Y.-Y. Yu, S.-W. Cao, Chin. J. Synthetic
Chem. 19 (2011) 349-351.
[36] G. Huang, S. Schramm, J. Heilmann, D. Biedermann,
V. Křen, M. Decker, Beilstein J. Org. Chem. 12 (2016)
662-669.
[37] I.F.F. Benzie, J.J. Strain, Anal. Biochem. 239 (1996)
70-76.
[38] R. Pulido, L. Bravo, F. Saura-Calixto, J. Agric. Food
Chem. 48 (2000) 3396-3402.
[39] G. Huang, B. Kling, F.H. Darras, J. Heilmann, M.
Decker, Eur. J. Med. Chem. 81 (2014) 15-21.
[40] F.H. Darras, B. Kling, J. Heilmann, M. Decker, ACS
Med. Chem. Lett. 3 (2012) 914-919.
[41] J.B. Davis, P. Maher, Brain Res. 652 (1994) 169-173.
[42] T.H. Murphy, M. Miyamoto, A. Sastre, R.L. Schnaar,
J.T. Coyle, Neuron 2 (1989) 1547-1558.
[43] S. Tan, M. Wood, P. Maher, J. Neurochem. 71 (1998)
95-105.
[44] S. Tan, D. Schubert, P. Maher, Curr. Top. Med. Chem.
1 (2001) 497-506.
[45] U. Saxena, Expert Opin. Ther. Pat. 16 (2012) 351-354.
[46] T. Wyss-Coray, J. Rogers, Cold Spring Harb.
Perspect. Med. 2 (2012) a006346.
[47] Y. Sagara, J. Vanhnasy, P. Maher, J. Neurochem. 90
(2004) 1144-1155.

ACCEPTED MANUSCRIPT
Graphical abstract:

ACCEPTED MANUSCRIPT
•
•
•
7-O-esters of the flavonolignan silibinin with natural acids were synthesized regioselectively.
Compounds show medium-term stability both in assay medium and plasma.
Esters with α,β-unsaturated acids show pronounced neuroprotection in HT-22 cells greatly
exceeding the effects of the components.
•
•
Cinnamic and ferulic acid esters exhibit overadditive effects regarding inhibition of microglial
activation, PC12 cell differentiation, in vitro ischemia as well as anti-aggregating abilities
against Aβ42 peptide and τ protein.
Ferulic acid esters with diastereomers silybin A and silybin B show a significant difference in
both neuroprotection and anti-aggregating capacities.