Semi-synthetic isoflavones as BACE-1 inhibitors against Alzheimer's disease

Article abstract of DOI:10.1016/j.bioorg.2019.03.034

BACE-1 is considered to be one of the targets for prevention and treatment of Alzheimer's disease (AD). We here report a novel class of semi-synthetic derivatives of prenylated isoflavones, obtained from the derivatization of natural flavonoids from Maclu

Full text of DOI:10.1016/j.bioorg.2019.03.034

Accepted Manuscript
Semi-synthetic isoflavones as BACE-1 inhibitors against Alzheimer’s disease
Giovanni Ribaudo, Paolo Coghi, Enrico Zanforlin, Betty Yuen Kwan Law, Yuki
Yu Jun Wu, Yu Han, Alena Congling Qiu, Brook Yuan Qing Qu, Giuseppe
Zagotto, Vincent Kam Wai Wong
PII:
S0045-2068(19)30146-4
DOI:
https://doi.org/10.1016/j.bioorg.2019.03.034
YBIOO 2860
Reference:
Bioorganic Chemistry
To appear in:
Received Date:
Revised Date:
Accepted Date:
30 January 2019
3 March 2019
14 March 2019
Please cite this article as: G. Ribaudo, P. Coghi, E. Zanforlin, B. Yuen Kwan Law, Y. Yu Jun Wu, Y. Han, A.
Congling Qiu, B. Yuan Qing Qu, G. Zagotto, V. Kam Wai Wong, Semi-synthetic isoflavones as BACE-1 inhibitors
against Alzheimer’s disease, Bioorganic Chemistry (2019), doi: https://doi.org/10.1016/j.bioorg.2019.03.034
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.

Graphical Abstract
To create your abstract, type over the instructions in the template box below.
Fonts or abstract dimensions should not be changed or altered.
Leave this area blank for abstract info.
Semi-synthetic isoflavones as BACE-1 inhibitors against Alzheimer’s disease
Giovanni Ribaudo[a] §, Paolo Coghi[b] §, Enrico Zanforlin[a] §, Betty Yuen Kwan Law[b], Yuki Yu Jun Wu[b], Yu
Han[b], Alena Congling Qiu[b], Brook, Yuan Qing Qu[b], Giuseppe Zagotto*[a] and Vincent Kam Wai Wong*[b]
aDr. G. Ribaudo, Dr. E. Zanforlin, Prof. G. Zagotto
Department of Pharmaceutical and Pharmacological Sciences, University of Padova, via Marzolo 5, 35131, Padova, Italy
E-mail: giuseppe.zagotto@unipd.it
bDr. P. Coghi, Miss Y. Y. J. Wu, Miss Y. Han, Dr. A. C. Qiu, Dr. B. Y. K. Law, Dr. V. K. W. Wong
State Key Laboratory of Quality Research in Chinese Medicine, Macau University of Science and Technology
Avenida Wai Long, Taipa, Macau, China
E-mail: bowaiwong@gmail.com
§ These authors contributed equally to this work

Bioorganic Chemistry
Semi-synthetic isoflavones as BACE-1 inhibitors against
Alzheimer’s disease
Giovanni Ribaudo[a] §, Paolo Coghi[b] §, Enrico Zanforlin[a] §, Betty Yuen Kwan Law[b], Yuki Yu Jun Wu[b],
Yu Han[b], Alena Congling Qiu[b], Brook Yuan Qing Qu[b], Giuseppe Zagotto*[a] and Vincent Kam Wai
Wong*[b]
aDr. G. Ribaudo, Dr. E. Zanforlin, Prof. G. Zagotto
Department of Pharmaceutical and Pharmacological Sciences, University of Padova, via Marzolo 5, 35131, Padova, Italy.
E-mail: giuseppe.zagotto@unipd.it
bDr. P. Coghi, Miss Y. Y. J. Wu, Miss Y. Han, Dr. A. C. Qiu, Dr. B. Y. K. Law, Dr. V. K. W. Wong
State Key Laboratory of Quality Research in Chinese Medicine, Macau University of Science and Technology
Avenida Wai Long, Taipa, Macau, China.
E-mail: bowaiwong@gmail.com
§ These authors contributed equally to this work.
ARTICLE INFO
ABSTRACT
Article history:
Received
Received in revised form
Accepted
BACE-1 is considered to be one of the targets for prevention and treatment of Alzheimer’s
disease (AD). We here report a novel class of semi-synthetic derivatives of prenylated
isoflavones, obtained from the derivatization of natural flavonoids from Maclura pomifera. In
vitro anti-AD effect of the synthesized compounds were evaluated via human recombinant
BACE-1 inhibition assay. Compound 7, 8 and 13 were found to be the most active candidates
which demonstrates good correlation between the computational docking and pharmacokinetic
predictions. Moreover, cytotoxic studies demonstrated that the compounds are not toxic against
normal and cancer cell lines. Among these three compounds, compound 7 enhance the activity
of P-glycoprotein (P-gp) on A549 cancer cells and increases the activity of P-gp ATPase with a
possible role on the efflux of amyloid-β across the blood- brain barrier. In conclusion, the
present findings may pave the way for the discovery of a novel class of compounds to prevent
and/or treat AD.
Available online
Keywords:
β-secretase(BACE-1)
isoflavones
P-gp
in silico molecular docking
Alzheimer’s disease
2009 Elsevier Ltd. All rights reserved.
therapeutic agents was pursed2f. BACE1 is the β-secretase
1. Introduction
enzyme cleaving the APP protein to release the C99 fragment
Alzheimer’s disease (AD) is a devastating, progressive
neurodegenerative disease that is associated with up to 80% of
the estimated 47 million cases of dementia worldwide, and is one
of the leading causes of death in the United States1a,b. While there
is no cure for AD1c, there are 5 prescriptions drugs currently
approved by food and drug administration (FDA) to treat its
symptoms. Among them Donepezil, Galantamine and
Rivastigmine are acetylcholinesterase (AChE) inhibitors (Fig. 1a),
while Memantine is N–methyl-D-aspartate receptor antagonist
(NMDA) (Fig 1b)1d.
that gives rise to various species of Aβ peptide during subsequent
γ-secretase cleavage2g.
Animal studies have indicated that prolonged treatment with
some BACE1 inhibitors may negatively affect spine formation
and density, hippocampal long-term potentiation, and cognition
in wild-type mice3. A significant number of BACE inhibitors
were abandoned because of poor tolerability and efficacy in man:
Verubecestat (MK-8931, Merck) 4a, Lanabecestat (AZD-3293,
LY3314814 Astra Zeneca and Eli Lilly) 4b, Atabecestat (JNJ-
54,861,911, Janssen)4c showed recently problems and their trials
were discontinued. Studies with derivative Elenbecestat4d and
others4e,f are ongoing and positive results on phase II/III were
recently published. Therefore although drug discovery targeting
the beta-amyloid remains controversial also recent studies4g on
multi-target directed ligands has provided insight and hope on
the use of natural products in targeting AD via BACE-beta
The majority of developed drugs in the last 20 years were
molecules interfering with aggregation of the β-amyloid (Aβ)
peptide. In 1991 Aβ accumulation was initially proposed2a-c as
the main event of AD, while amyloid cascade hypothesis was
proposed one year later2d and in course of years has evolved2e.
BACE-1 is emerged as an important drug target for reducing Aβ
levels in the AD and the development of BACE-1 inhibitors as

amyloid mechanism. On the other hand, P-glycoprotein (P-gp)
is a protein involved in the active transport of molecules across
biological membranes, and it has been shown that alterations of
its expression and/or activity are related to the onset of
neurodegenerative disorders. Indeed, it has been reported that
up-regulation of this ATP-dependent dependent efflux pump is
necessary for a decrease of Aβ intracellular accumulation,5
suggesting the possible therapeutic approach for AD by
activation of P-gp activity.
followed by an opportune separation procedure to obtain 1 and 2,
while chloroform was used to isolate 3 and 4. Despite the very
close structural similarity among these molecules, ‘diagnostic
1
peaks’ in the H NMR spectrum and NOESY NMR experiment
allow to unambiguously distinguish the compounds and to
identify the correct structures 9,11
.
We performed the extraction of 1 – 4 according to a
previously described procedure that allows the acquisition of the
two couples of isomers in comparable yields12.The natural and
semi-synthetic derivatives (Fig. 2) that were prepared starting
1
from these four isoflavones were fully characterized by H and
13C NMR and high-resolution mass spectrometry. The purity
profile was assessed by HPLC.
Figure 1. Structures of FDA-approved treatments for Alzheimer’s (a – b)
and derivatives BACE-1 inhibitors ongoing and recently discontinued(c).
Recently, natural and semi-synthetic compounds captivated
the interest of medicinal chemists, as they may represent
promising and alternative “lead compounds” to be developed as
novel therapeutic remedy against CNS-related diseases6. For
example, flavonoids are well known natural compounds
possessing a broad range of pharmacological neuroprotective
properties related to AD 7a, such as supressive effect on AChE
activities7b, Aβ fibril formation 7c, and H2O2 – induced ROS
formation7d. Recent studies have demonstrated that
isoflavonoids such as genistein and biochanin A, are able to
inhibit the activity of BACE-18.
Figure 2. Structures of natural compounds 1-4 and derivatives synthesized
and investigated 5 – 16.
Isoosajin (14) and isopomiferin (16) were synthesized by
adapting the procedure reported by Wolfrom et al. which heat the
natural compounds in acetic and sulfuric acid13. These
compounds were synthesized with the aim of exploring the effect
of an increased rigidity of the scaffold, together with a reduced
flexibility and mobility of the side chains. Nevertheless, such
modifications on osajin and pomiferin structures led to the loss of
one hydrogen bond donor site on the molecules. The synthesized
compounds were also used as intermediates for the preparation of
the derivatives described below.
We have previously described the biochemical properties of
derivatives derived from Pomiferin 1 and Osajin 2, which have
been confirmed as a phosphodiesterase inhibitors9. These two
isoflavones isolated from Maclura Pomifera (Osage orange)
were also found to exhibit anti-cancer10a, anti-bacterial10b and
anti-diabetic properties10c. In this study, we further investigated
the potential application of using semi-synthetic isoflavones as
BACE-1 inhibitors and P-gp modulators for modulating AD. We
performed accurate and wide in silico studies on this class of
molecules focusing on the computational definition of their
physical and chemical properties, and their possible mode of
interaction with the target enzyme BACE-1 and P-gp. These
preliminary results were validated by in vitro test performed for
the evaluation of the efficacy of these compounds. By using
cytotoxicity assay, we further confirmed that these compounds
are not toxic towards both normal and cancerous cells. Most
importantly, among all derivatives which showed promising
inhibitory activity towards BACE-1 with the activation on the
activity of P-gp, compound 7 is the most promising candidate.
The
preparation
of
methanesulfonate
and
4-
methylbenzenesulfonate derivatives was carried out according to
a procedure previously described by Wolfrom et al.13 However,
these authors only proposed the preparation of the 4-
methylbenzenesulfonate derivatives of pomiferin and osajin and
their cyclized isoosajin and isopomiferin analogues. We therefore
synthesized these molecules by reacting other structurally related
isoflavones under similar experimental conditions (Supporting
information, Scheme 1 and 2). Moreover, we expanded the
chemical class by introducing the methanesulfonate moiety
(Supp. information, Scheme 1 - 4). The introduction of the -SO2R
moiety was carried out to explore the effects of extending the
bulkiness of the molecule in the phenyl region, and to shade the
hydroxyl groups, and thereby introduced a novel hydrogen bond
acceptor 14. This modification is particularly useful because these
new functional groups interact strongly with the positive charged
amino-acids in the region immediately closed to the active site of
BACE-1 as revealed by docking studies. The preparation of such
compounds were carried out by reacting the opportune osajin or
pomiferin derivative with methanesulfonyl chloride or 4-
methylbenzenesulfonyl chloride in pyridine, according to the
procedure described in the experimental section. The newly
synthesized compounds are shown in Fig. 2.
2. Result and discussion
2.1. Chemistry.
Auriculasin (3) and Scandenone (4) are defined as ‘chemically
constitutional isomers’ of pomiferin (1) and osajin (2), and are
the well-known isoflavones that can be isolated from M.
pomifera fruits11. In fact, these fruits are rich sources of both sets
of isomers, which can be extracted with a Soxhlet apparatus.
Ethyl acetate or diethyl ether were used as the extraction solvent,

2.2. In vitro BACE-1 inhibitory activity Assay
Table 1. Cytotoxicity of 1-16 against against human liver HepG2, lung A549
cancer cell lines, the normal human hepatocytes, LO2 and normal lung epithelial
cells, BEAS-2B.
With the recognition of BACE-1 proteases as a key player in
the cleavage of amyloid precursor protein (APP) to produce toxic
Aβ peptide, the development of BACE-1 inhibitors has become
one of the most promising approaches in AD drug discovery. By
adopting a fluorometric-based method utilizing a secretase-
specific peptide conjugated to the reporter molecules (EDANS
and DABCYL), uncleaved peptide showed the basal fluorescent
emissions from EDANS, due to the quenching by the physical
interaction of the DABCYL. In contrast, cleavage of the peptide
by β- secretase leads to the physical separation of EDANS and
DABCYL, resulting in the emission of fluorescent signal (Ex/
Em = 345/500 nm), which is proportional to the β-secretase
enzymatic activity. As shown in Figure 3, compounds 7, 8, 9, 10
and 13 with their IC50 value in BEAS-2B normal cells showed
the most potent inhibitory effect on β-secretase enzymatic
activity, suggesting their potential inhibitory effect on Aβ peptide
level.
BEAS-2B
cells
IC50 [µM]
A549
cells
IC50 [µM]
LO2
cells
IC50
HEPG2
cells
IC50
Compounds
[µM]
[µM]
Donezepila
n/a
>200
n/a
n/a
1
9.8
50.09
48.6
n/a
15.8
70.01
n/a
18.9
31.4
n/a
2
16.9
13.4
18.9
>100
>100
>100
>100
92.6
>100
>100
>100
>100
18.2
>100
14.1
3
4
36.3
>100
>100
>100
>100
18.4
>100
>100
>100
>100
21.2
>100
>100
22.8
>100
86.1
>100
>100
42.6
>100
>100
>100
>100
11.3
>100
11.8
41.2
>100
>100
>100
>100
88.1
>100
>100
>100
>100
13.1
>100
28
5
6
7
8
9
10
11
12
13
14
15
16
Figure 3. β-secretase activity of compounds 4-16
2.3. Cytotoxic effect in human cancerous and normal cells
[a] ref. 27.
To support the above reported predictions, cell viability assay
was performed to study the cytotoxic effect of these compounds
in different cells lines.
2.4. Activity of isoflavones derivatives 4 – 16 on P-gp
The pathogenesis of AD may be caused by an imbalance
between β-amyloid peptide (Aβ) production and clearance. Some
studies reported that β-amyloid is also a substrate for P-gp, and in
vitro blocking of P-gp functions decreases its clearance across the
BBB5. For instance, rivastigmine enhanced clearance of Aβ
through up-regulation of P-gp15. Therefore, targeting β-amyloid
clearance by P-gp stimulation could be a useful strategy to
prevent AD progression16. The biological evaluation of
compounds 7, 8 and 13 was carried out in A549 tumor cell lines16.
In this assay, rhodamine-123, a fluorescent probe, is used as the
P-gp or MRP1 substrate. Recently17 we have demonstrated that
isoflavones pomiferin and isopomiferin are strong P-gp inhibitors
(10 μM) in comparison with VP in multidrug resistant cancer
cells (Supporting information, Table 7). At low concentration
(100 μM), we found slightly inhibitory effects of 8 and 13 on P-
gp of A549 cancer cells, while compound 7 showed significant
decreasing of rhodamine different towards control (Supporting
information, Table 8).
The cytotoxicity of the analysed compounds was examined in
normal human hepatocytes, LO2 and normal lung epithelial cells,
BEAS-2B. Among the 16 investigated compounds, 62% and 56%
of the compounds showed no cytotoxicity in both normal liver
LO2 and lung BEAS-2B cell lines (Table 1, Supporting
information Table 2 and 3). Moreover, compounds were also
examined on human liver and lung cancer cell lines (HepG2 and
A549 (Table 1, Supporting information Table 5 and 6). Results
confirmed that compounds 7, 8 and 13 showed no cytotoxicity on
both cancer cells lines while compounds 1 and 2 and respective
isomers 14 and 16 revealed moderate cytotoxicity (IC50 13-32
μM) on HepG2 cell lines. Compound 9 present better selectivity
for lung cancer A549 cell lines (5 times).
Derivatives 7, 8, 10 and 13 with potent inhibitory effect on β-
secretase enzymatic activity showed no toxicity in all tested cell
lines. Natural compounds 1 and 2 and respective isomers 14 and
16 were also evaluated for their cytotoxicity against normal
human lung fibroblasts CCD19Lu showed moderate cytotoxicity
(IC50 22-64 μM). (Supporting information Table 4).
We then further evaluated the effects of compound 7 on P-gp
ATPase activity. Drug stimulated P-gp ATPase activity was
estimated by P-gp-Glo assay system (Promega, Madison, WI)
and the results were shown in Fig. 4 and Supporting information
table 9. This method relies on the ATP dependence light-

generating reaction of firefly luciferase where ATP consumption
is detected as a decrease in luminescence (Fig. 4a). The greater
extent on the drop of luminescence signal indicates the stronger
the P-gp activity. Here, sodium orthovanadate Na3VO4 was used
as a P-gp ATPase inhibitor, whereas verapamil (200 μM, VP)
was used as a P-gp stimulator. While compounds 7 (tested at 50
μM, 100 μM and same concentration of VP) with luminescence
signal lower than the control (NT) is considered as a P-gp
stimulator, compounds with luminescence signal higher than the
NT are considered as P-gp inhibitors. Compounds which are not
significantly different from the NT are neither P-gp stimulator
nor inhibitor (Supporting information table 9). Compound 7 was
demonstrated as a novel P-gp stimulator (Fig. 4b, at 100 μM,
Supp.info table 8 and 9) that might influence efflux of β amyloid
across the BBB.
effects of 7 on efflux of Aβ by P-gp is required16,19 in coming
future.
Figure 5. Docking pose of Verapamil (red) and compound 7 (blue) in binding
site of P-gp
2.5. Molecular Docking Studies and In Silico Physicochemical
Prediction.
In computational study, we first focused on the
pharmacokinetic properties of the new compounds, and in
particular on those parameters, for example, logP, Topological
polar surface area (TPSA) and lipophilicity in general, that are
fundamental for crossing the Blood Brain Barrier (BBB). The
“drug likeness” of this selected set of compounds was therefore
investigated and predicted for their pharmacokinetic properties
by
using
Molinspiration
Cheminformatics
software
(http://www.molinspiration.com). We considered the logP value,
which is also known to reflect the key physical properties of
compounds such as solubility, lipophilicity and bioavailability.
TPSA, volume and molecular weight, which are respectively
expressed in Å2, Å3 and Da, provide information on the
bulkiness of the molecule and the properties of its surface. The
number of hydrogen bond donors (nONH) and hydrogen bond
acceptors (nON) described the specificity of the recognition
between ligand and substrate, and helped to define the propensity
of the molecule to generate a-specific interactions with other
substrates. The number of rotatable bonds defined the flexibility
of the compound. These parameters influence membrane
diffusion, transport properties such as blood-barrier crossing and,
in general, bioavailability and the percentage of absorption. As
reported by Pajouhesh,20 a successful CNS drug or new chemical
entity should fulfill certain parameters, such as: MW < 450Da,
clogP < 5, nON < 7, nONH < 3, number of rotatable bonds < 8.
The analysis of the pharmacokinetic properties on our
synthesized compounds highlighted that while some of them only
partially met the above requirements, the most promising
candidates satisfied with these conditions more closely
(Supporting information Table 1).
Figure 4. Stimulation of P-gp ATPase activity by Verapamil (VP) and 7.
Verapamil, which is supplied as a positive control substrate with the Pgp-Glo™
Assay System. Untreated (NT), 100µM Na3VO4- and 200µM Verapamil-treated
Pgp reactions were performed according to the protocol described in
experimental section. Luminescence was read on a SpectraMax Paradigm Multi-
Mode Microplate Reader (Panel a). The decrease in luminescence of NT
samples compared to samples plus Na3VO4 (ΔRLUbasal) represents basal Pgp
ATPase activity. The decrease in luminescence of Verapamil-treated samples
(ΔRLUTC, TC test compound) represents Verapamil-stimulated Pgp ATPase
activity. ΔRLUbasal and ΔRLUTC were replotted (Panel b) to illustrate the
stimulation of Pgp ATPase activity by Verapamil and 7 (see Supp. inform. Table
9B).
With the presence of both compounds 7 and VP, no further
enhancement of P-gp ATPase activity confirmed the docking
results which suggested both compound 7 and VP are binding
and competing for same binding domain of P-glycoprotein (Fig
5). Moreover,18 derivative indolizin sulfone SR3357, containing
benzenesulfonyl core similarly to 7, was found interacting with a
multiple binding sites on P-gp which are allosterically linked and
have differential effect on the ATPase activity of the protein.
Therefore, further research work to evaluate the stimulating
These data help to identify the compounds with good oral
bioavailability and penetration rate to the central nervous system
(CNS), and provide the crucial parameter needed for the design
of BACE-1 inhibitors.
However, it was observed that all extracted natural
compounds (1 – 4), demonstrated negative score and hence were
not considered as a BBB penetrant. All compounds containing
sulfonic portions showed value above the threshold of 0.02,
demonstrating the introduction of –SO2R moiety may help to

Figure 6: A. Superimposition of all the best poses of the active molecules that confirm the hypothesis of a common site of action (for the singular poses of the
most promising compounds see the Supporting information). B. Average values, of three different experiments, of the energies of interactions between BACE-1
and each tested compound.
increase the penetration rate of compound to BBB. Previous
studies reported the addition of sulphates for dopamine21a and
steroids21b,c could increase their BBB permeation as sulphate
group is known substrate of specific transporters. While the
Figure 7: 2D interaction plot of the compound 7 within the active site of
active compounds 7, 8, 9, 10 and 13 (Supporting information
BACE-1.
table 1) possessed scores ranging from 0.055 in compound 9 to a
These interactions were maintained by almost all the active
a maximum value of 0.105 in compound 13, compounds 11 and
compounds. Other hydrophobic interactions occurred between
15 showed a significant high score in BBB permeation as
BACE-1 and the active compounds, such as the interactions with
revealed by in silico study, although with no inhibitory activity
Phe108, Trp115, Ile118 and Leu30 residues.
on BACE-1 observed in vitro (Fig 3).
The common pattern of interaction, together with the binding
energies of the active compounds, are correlated with their
Accordingly, an accurate docking study has been performed to
understand the mode of interaction of our compounds with
biological properties. Indeed, the energies of interaction,
BACE-1 and P-gp.
expressed in Kcal/mol, fit very well with the decrease of activity
Following a blind docking technique, the preliminary virtual
screening has shown that all tested compounds were able to
interact selectively and strongly with the active site of BACE-1,
as showed in Figure 6B. Based on the results of this preliminary
in silico study, compounds 7, 8, 11, 14 and 15 appeared to be the
most promising candidates. Comparing these results with the in
vitro activities, a direct correlation rises with compounds 7 and 8
being identified as the most potent inhibitors of BACE-1.
of the enzyme. This result further confirmed the relevance of this
set of compounds as inhibitors of BACE-1. Moreover,
considering the obtained energy values for the tested compounds,
they were comparable or even better than the calculated energy
for the oxazine-based compound22a which has been co-crystalized
with BACE-1 (PDB ID: 5CLM) (Supp. information Fig 1a).
Nevertheless, it must be pointed out that the predicted
interactions of the isoflavones seem to exclude a direct
enrollment of the catalytic dyad formed by Asp 32 and Asp 228.
Thus, further modification of the proposed scaffold with the aim
of directly interfering with these amino acids could represent an
interesting improvement in term of drug specificity and potency.
Together with all the compounds which showed activities in
in vitro tests, compounds 9, 10 and 13 were further investigated
by using further in silico study with higher accuracy, and with the
focus limiting into the active site of the enzyme and the
surrounding area (Supp. information Fig. 2). Interestingly, this
experiment has demonstrated a defined pattern of interaction
between those biologically active compounds. As reported in the
2D interactions plots in Fig.7 and in the Supp. information Fig.
2a-e, the active compounds formed several H-bonds, especially
with the Arg235 and Asn233 residues, exploiting their sulfonic
portions.
A second set of in silico experiments have been conducted to
evaluate and identify the possible site of interaction within P-gp
and the allosteric modulators.22b Also, blind docking has been
adopted to evaluate the most promising compounds (7, 8, 13).
Verapamil (VP) was used as a positive control drug for
validating the possible site of action of the tested compounds. All
these compounds, excluding derivative 13, interacted with the
same pocket within the P-gp with their energies of interactions
correlated positively with the obtained biological results
(Supplementary information Fig 1b). As reported in the 2D
interactions plots in the Supp. information Fig. 3a-e, the active
compounds presented different modes of interaction with P-gp.
As shown in Figure 5, the pose of VP within P-gp strongly
overlapped to that of compound 7, which is the most promising
candidate from all of the tested compounds. Compound 7
exhibits two hydrogen bonds with the Tyr303 and the Gly222
exploiting their sulfonic portion, generating the important π-π
stacking with the amino acids Tyr306 and Phe299, which

4.3.1 Characterization of 7-(3,4-dihydroxyphenyl)-9-hydroxy-
2,2-dimethyl-10-(3-methyl-2-butenyl)-1,5-dioxa-2H-
phenanthren-8-one (1, pomiferin)
positions were also involved in binding pose of verapamil.
Several other hydrophobic interactions contributed to the binding
energy of our leader compound (Supp. information Fig 1b).
3. Conclusion
Yield: 0.7 g from 200 g of dried M. pomifera fruits. δH (400
MHz, CDCl3): 12.97 (1H, s, OH), 7.89 (1H, s, C=CH), 7.04 (1H,
d, J 1.1 Hz, Ar-H), 6.84 (1H, d, J 7.1 Hz, C=CH), 6.71 (1H, dd, J
7.1 Hz, Ar-H), 5.56 (1H, d, J 10 Hz, C=CH), 5.24 (1H, m,
C=CH), 3.36 (2H, d, J 7.0 Hz, CH2), 1.82 (3H, s, CH3), 1.69 (3H,
s, CH3), 1.50 (6H, s, CH3). δC (100 MHz, CDCl3): 181.2, 159.3,
157.4, 152.6, 150.4, 144.8, 144.1, 131.6, 127.4, 123.8, 122.8,
121.9, 121.8, 116.6, 115.6, 115.2, 113.2, 105.6, 100.9, 77.8, 28.3,
26.2, 21.6, 18.1. HMRS (ESI) found 421.1653, calc. 421.1646
Anal. found C 71.5, H 5.8. Calc. for C25H24O5: C 71.4, H 5.8.
The prenylated isoflavones obtained from the derivatization of
natural flavonoids from Maclura pomifera showed promising in
vitro and in silico results in terms of BACE-1 inhibition and P-gp
stimulation. The compounds were evaluated for their oral
bioavailability, blood brain barrier permeability and drug
likeliness in general (folding, polarity, and molecular size).
The present study demonstrated that prenylated isoflavones in
particular 7, 8 and 13 are active on BACE-1. Furthermore, our
result provided important mechanistic insights into the binding
mechanism to the enzyme. Moreover compound 7 showed
stimulating effect on efflux of rhodamine and P-gp ATPase
activity. The current study provided for the first time the
important preliminary information for identifying and
characterizing prenylated isoflavones as dual targets for the
treatment and/or prevention of AD.
4.3.2 Characterization of 9-hydroxy-7-(p-hydroxyphenyl)-2,2-
dimethyl-10-(3-methyl-2-butenyl)-1,5-dioxa-2H-phenanthren-8-
one (2, osajin)
Yield: 4.0 g from 200 g of dried M. pomifera fruits. δH (400
MHz, CDCl3): 13.14 (1H, s, OH), 7.88 (1H, s, C=CH), 7.39 (2H,
m, Ar-H), 6.89 (2H, m, J 8.1 Hz, ArH), 6.70 (1H, d, J 10 Hz,
C=CH), 5.59 (1H, d, J 10 Hz, C=CH), 5.26 (1H, m, J 7.3 Hz,
C=CH), 5 (1H, s, OH), 3.37 (2H, d, J 7.3 Hz, CH2), 1.83 (3H, s,
CH3), 1.70 (3H, s, CH3), 1.50 (6H, s, CH3). δC (100 MHz,
CDCl3): 181.1, 159.3, 157.4, 155.0, 152.4, 150.6, 131.7,
130.5,127.3, 123.7, 123.0, 122.0, 115.8, 115.1, 113.0, 105.6,
100.8, 77.9, 28.3, 26.0, 21.4, 18.0. HMRS (ESI) found 405.1669,
calc. 406.1697 Anal. found C 74.1, H 6. Calc. for C25H24O5: C
74.2, H 6.
4. Experimental
4.1. General
Human normal hepatocytes, LO2, lung epithelial cells, BEAS-
2B and human normal fibroblasts, CCD19Lu were all purchased
from ATCC. Cells were cultured in RPMI-1640 medium
supplemented with 10% fetal bovine serum and antibiotics
penicillin (50 U/ml) and streptomycin (50μg/ml; Invitrogen,
U.K.). All cells were incubated at 37°C in a 5% humidified CO2
incubator).
4.3.3 Characterization of 7-(3,4-dihydroxyphenyl)-5-hydroxy-
2,2-dimethyl-10-(3-methylbut-2-enyl)pyrano[3,2-g]chromen-6-
one (3, auriculasin)
4.2. Chemistry
Yield: 0.8 g from 200 g of dried M. pomifera fruits. δH (400
MHz, CDCl3): 13.46 (1H, s, OH), 7.80 (1H, s, C=CH), 6.94 (1H,
d, J 1.1 Hz, Ar-H), 6.73 (1H, d, J 7.1 Hz, C=CH), 6.89 (1H, dd, J
7.1 Hz, Ar-H), 6.65 (1H, dd, J 10.0 Hz, C=CH), 5.67 (1H, d, J 10
Hz, C=CH), 5.20 (1H, m, C=CH), 5.03 (1H, s, OH), 3.32 (2H, d,
J 7.0 Hz, CH2), 1.80 (3H, s, CH3), 1.65 (3H, s, CH3), 1.48 (6H, s,
CH3). δC (100 MHz, CDCl3): 182.0, 156.2, 156.0, 154.9, 150.1,
144.6, 144.1, 131.6, 127.1, 123.8, 122.2, 121.9, 121.0, 116.6,
115.9, 115.6, 113.2, 108.4, 100.9, 77.8, 28.3, 26.7, 22.1, 17.2.
HMRS (ESI) found 421.1652, calc. 421.1646 Anal. found C 71.5,
H 5.8. Calc. for C25H24O5: C 71.4, H 5.8.
Commercially available chemicals were purchased from
1
Aldrich unless otherwise stated. H and 13C {1H} NMR spectra
were recorded on a Bruker Avance III 400 MHz spectrometer
and on a Bruker AMX 300 MHz spectrometer. All spectra were
recorded at room temperature, the solvent for each spectrum is
given in parentheses. Chemical shifts are reported in ppm and are
relative to TMS internally referenced to the residual solvent peak.
Datasets were edited with Bruker TopSpin suite and iNMR. The
multiplicity of signals is reported as singlet(s), doublet (d), triplet
(t), quartet (q), multiplet (m), broad (br) or a combination of any
of these. High resolution mass spectra (HRMS) were recorded on
a ESI–TOF Mariner from Applied Biosystems using electrospray
ionization (ESI). The purity profile of the compounds was
assayed by HPLC using a Varian Pro-Star system equipped with
a Biorad 1706 UV–VIS detector and an Agilent C-18 column (5
μm, 4.6 × 150 mm). An appropriate ratio of water (A) and
acetonitrile (B) was used as mobile phase with an overall flow
rate of 1 mL min− 1. The general method for the analyses is
reported here: 0 min (90% A–10% B), 5 min (90% A–10% B),
25 min (10% A–90% B), 30 min (90% A–10% B), and 32 min
(90% A–10% B). The purity of all compounds was ≥96%, unless
otherwise stated (254 nm).
4.3.4.Characterization of 5-hydroxy-7-(4-hydroxyphenyl)-2,2-
dimethyl-10-(3-methylbut-2-enyl)pyrano[3,2-g]chromen-6-one
(4, Scandenone)
Yield: 3.6 g from 200 g of dried M. pomifera fruits. δH (400
MHz, CDCl3): 7.89 (1H, s, C=CH), 7.35 (2H, m, Ar-H), 6.84 (2H,
m, J 8.1 Hz, ArH), 6.74 (1H, d, J 10 Hz, C=CH), 5.63 (1H, d, J
10 Hz, C=CH), 5.17 (1H, m, J 7.3 Hz, C=CH), 5 (1H, s, OH),
3.40 (2H, d, J 7.3 Hz, CH2), 1.81 (3H, s, CH3), 1.68 (3H, s, CH3),
1.47 (6H, s, CH3). δC (100 MHz, CDCl3): 181.4, 159.2, 157.4,
155.9, 152.6, 150.6, 131.2, 130.5,127.1, 123.7, 123.0, 121.9,
115.8, 115.6, 113.0, 105.6, 100.8, 77.9, 28.3, 25.7, 21.3, 17.9.
HMRS (ESI) found 405.1670, calc. 405.1697 Anal. found C 74.1,
H 6. Calc. for C25H24O5: C 74.2, H 6.
4.3. General procedure for isolation of pomiferin (1) and osajin
(2), auriculasin (3) and scandenone (4)
Compounds were extracted with a Soxhlet apparatus from .
Ethyl acetate or diethyl ether were used as the extraction solvent,
followed by an opportune separation procedure to obtain 1 and 2,
while chloroform was used to isolate 3 and 4.
4.4. General procedure for the preparation of methanesulfonate
derivatives

A round bottom flask was charged with the opportune
isoflavone (200 mg) and 2 mL of pyridine. A solution of 0.6 mL
of methanesulfonyl chloride in pyridine (3 mL) was added and
the resulting mixture was stirred at room temperature. The
reaction was monitored by TLC (hexane/ethyl acetate 1:1) and,
after completion, it was quenched by pouring the mixture in cold
water. The resulting solid was collected by filtration and
recrystallized from ethanol.
110.2, 109.8, 108.9, 108.2, 107.2, 105.4, 102.1, 77.1, 29.1, 28.5,
25.2, 21.6, 17.9. HMRS (ESI) found 577.1642, calc. 577.1124.
Anal. found C 56.2, H 4.9, S 11.1. Calc. for C27H28O10S2: C 56.2,
H 4.9, S 11.1%.
4.4.5 Characterization of 4-(2,2,10,10-tetramethyl-8-oxo-
8,10,11,12-tetrahydro-2H-dipyrano[2,3-f:2',3'-h]chromen-7-yl)-
1,2-phenylene dimethanesulfonate (13)
4.4.1. Characterization of 4-(5-hydroxy-8,8-dimethyl-6-(3-
methylbut-2-en-1-yl)-4-oxo-4,8-dihydropyrano[2,3-f]chromen-3-
yl)-1,2-phenylene dimethanesulfonate (5)
Yield: 71%. δH (400 MHz, CDCl3): 7.84 (1H, s, C=CH), 7.73
(1H, d, J 1.8, Ar-H), 7.59 (1H, d, J 8.4 Hz, Ar-H), 7.53 (1H, d, J
8.4 Hz, Ar-H), 6.75 (1H, d, J 10.0 Hz, C=CH), 5.60 (1H, d, J
10.0 Hz, C=CH), 3.28 (3H, s, CH3), 3.27 (3H, s, CH3),2.69 (2H,
m, CH2), 1.86 (2H, m, CH2), 1.52 (6H, s, CH3), 1.45 (6H, s, CH3).
δC (100 MHz, CDCl3): 182.1, 159.2, 156.1, 152.0, 150.2, 144.5,
140.2, 137.5, 128.1, 119.0, 113.1, 110.0, 108.1, 108.0, 103.1,
102.2, 89.1, 83.5, 38.1, 37.6, 33.3, 29.4, 27.1, 26.1. HMRS (ESI)
found 729.1745, calc. 729.1758. Anal. found C 63.1, H 5.1, S 8.8.
Calc. for C39H39O10S2: C 63.1, H 5.0, S 8.8%.
Yield: 71%. δH (400 MHz, CDCl3): 12.88 (1H, s, OH), 7.98
(1H, s, C=CH), 7.74 (1H, m, Ar-H), 7.57 (2H, m, Ar-H), 6.70
(1H, d, J 9.9 Hz, C=CH), 5.62 (1H, d, J 9.9 Hz, C=CH), 5.23 (1H,
m, C=CH), 3.36 (2H, d, J 7.1 Hz, CH2), 3.31 (3H, s, CH3), 3.30
(3H, s, CH3), 1.83 (3H, s, CH3), 1.70 (3H, s, CH3), 1.51 (6H, s,
CH3). δC (100 MHz, CDCl3): 182.1, 161, 159.2, 154.6, 150.9,
133.5, 132.3, 132.2, 129.1,124.3, 124, 123.5, 116.5, 115, 107.2,
102.6, 79.7, 78.4, 39.2, 29.9, 27.5, 23, 19.7. HMRS (ESI) found
577.1642, calc. 577.1124 Anal. found C 56.1, H 4.9, S 11.1. Calc.
for C27H28O10S2: C 56.2, H 4.9, S 11.1%.
4.5. General procedure for the preparation of 4-
methylbenzenesulfonate derivatives
A round bottom flask was charged with the opportune
isoflavone (100 mg) and 1 mL of pyridine. A solution of 400 mg
of 4-methylbenzenesulfonyl chloride in pyridine (2 mL) was
added and the resulting mixture was stirred at room temperature.
The reaction was monitored by TLC (hexane/ethyl acetate 1:1)
and, after completion, it was quenched by pouring the mixture in
cold water. The resulting solid was collected by filtration and
recrystallized from ethanol.
4.4.2. Characterization of 4-(5-hydroxy-8,8-dimethyl-6-(3-
methylbut-2-en-1-yl)-4-oxo-4,8-dihydropyrano[2,3-f]chromen-3-
yl)phenyl methanesulfonate (6)
Yield: 69%. δH (400 MHz, CDCl3): 13.02 (1H, s, OH), 7.93
(1H, s, C=CH), 7.60 (2H, d, J 8.8 Hz, Ar-H), 7.38 (2H, d, J 8.8
Hz, Ar-H), 6.71 (1H, d, J 9.9 Hz, C=CH), 5.61 (1H, d, J 9.9 Hz,
C=CH), 5.25 (1H, m, C=CH), 3.36 (2H, d, J 7.1 Hz, CH2), 3.20
(3H, s, CH3), 1.83 (3H, s, CH3), 1.71 (3H, s, CH3), 1.50 (6H, s,
CH3). δC (100 MHz, CDCl3): 180.3, 159.2, 157.4, 152.8, 150.4,
149.1, 131.7, 130.6, 130.4, 127.4, 122.5, 122.1, 121.7, 114.7,
113.1, 105.4, 100.8, 78, 37.4, 28.1, 25.8, 21.3, 17.9. HMRS
(ESI) found 483.1695, calc. 483.1399. Anal. found C 64.7, H 5.4,
S 6.5. Calc. for C26H26O7S C 64.7, H 5.4, S 6.6%.
4.5.1. Characterization of 4-(5-hydroxy-8,8-dimethyl-6-(3-
methylbut-2-en-1-yl)-4-oxo-4,8-dihydropyrano[2,3-f]chromen-3-
yl)phenyl 4-methylbenzenesulfonate (7)
Yield 68%. δH (400 MHz, CDCl3): 12.98 (1H, s, OH), 7.89
(1H, s, C=CH), 7.77 (2H, m, Ar-H), 7.49 (2H, m, Ar-H), 7.35
(2H, d, J 7.9 Hz, Ar-H), 7.08 (2H, m, Ar-H), 6.69 (1H, d, J 9.9
Hz, C=CH), 5.60 (1H, d, J 9.9 Hz, C=CH), 5.23 (1H, m, C=CH),
3.36 (2H, m, CH2), 2.49 (3H, s, CH3), 1.83 (3H, s, CH3), 1.70
(3H, s, CH3), 1.50 (6H, s, CH3). δC (100 MHz, CDCl3): 180.2,
159.3, 152.8, 145.4,130.1, 130, 129.8, 128.5, 127.3, 122.4, 121.7,
114.8, 28.1, 25.7, 21.7, 17.9. HMRS (ESI) found 559.2178, calc.
559.1712. Anal. found C 68.9, H 5.4, S 5.6. Calc. for C32H30O7S
C 68.8, H 5.4, S 5.7%.
4.4.3 Characterization of 4-(5-hydroxy-8,8-dimethyl-10-(3-
methylbut-2-en-1-yl)-4-oxo-4,8-dihydropyrano[3,2-g]chromen-3-
yl)phenyl methanesulfonate (9)
Yield: 55%. δH (400 MHz, CDCl3): 13.00 (1H, s, OH), 7.93
(1H, s, C=CH), 7.59 (2H, d, J 8.8 Hz, Ar-H), 7.40 (2H, d, J 8.8
Hz, Ar-H), 6.70 (1H, d, J 9.9 Hz, C=CH), 5.61 (1H, d, J 9.9 Hz,
C=CH), 5.25 (1H, m, C=CH), 3.38 (2H, m, CH2), 3.20 (3H, s,
CH3), 1.83 (3H, s, CH3), 1.70 (3H, s, CH3), 1.50 (6H, s, CH3). δC
(100 MHz, CDCl3): 180.2, 156.7, 155.5, 154.2, 154.0, 153.1,
131.8, 131.2, 129.1, 124.1, 122.2, 120.9, 116.1, 108.1, 103.1,
78.2, 37.1, 29.1, 28.1, 25.9, 21.3, 18.0. HMRS (ESI) found
483.1695, calc. 483.1399. Anal. found C 64.8, H 5.4, S, 6.5. Calc.
for C26H26O7S: C 64.7, H 5.4, S, 6.6%.
4.5.2. Characterization of 4-(5-hydroxy-8,8-dimethyl-6-(3-
methylbut-2-en-1-yl)-4-oxo-4,8-dihydropyrano[2,3-f]chromen-3-
yl)-1,2-phenylene dimethanesulfonate (8)
Yield: 58%. δH (400 MHz, CDCl3): 12.96 (1H, s, OH), 7.91
(1H, s, C=CH), 7.70 (4H, m, J 1.8, Ar-H), 7.46 (1H, m, Ar-H),
7.46 (1H, m, Ar-H), 7.35 (1H, m, Ar-H), 7.33 (4H, m, Ar-H),
6.69 (1H, d, J 9.9 Hz, C=CH), 5.61 (1H, d, J 9.9 Hz, C=CH),
5.23 (1H, m, C=CH), 3,36 (2H, d, J 7.1 Hz, CH2), 2.48 (6H, s,
CH3), 1.83 (3H, s, CH3), 1.70 (3H, s, CH3), 1.51 (6H, s, CH3). δC
(100 MHz, CDCl3): 180.8, 159.2, 157.5, 145.7, 141, 132.1, 131.7,
129.8, 128.6, 128.5, 125, 124.3, 122.1, 121.7, 28.8, 28.1, 25.9,
21.8. HMRS (ESI) found 729.2207, calc. 729.1750. Anal. found
C 64.2, H 5.0, S 8.8. Calc. for C39H36O10S2: C 64.3, H 5.0, S
8.8%.
4.4.4. Characterization of 4-(5-hydroxy-8,8-dimethyl-10-(3-
methylbut-2-en-1-yl)-4-oxo-4,8-dihydropyrano[3,2-g]chromen-3-
yl)-1,2-phenylene dimethanesulfonate (12)
Yield: 67%. δH (400 MHz, CDCl3): 12.77 (1H, s, OH), 7.98
(1H, s, C=CH), 7.74 (1H, m, Ar-H), 7.58 (2H, m, Ar-H), 6.72
(1H, d, J 9.9 Hz, C=CH), 5.64 (1H, d, J 9.9 Hz, C=CH), 5.26 (1H,
m, C=CH), 3.37(2H, d, J 7.1 Hz, CH2), 3.31 (3H, s, CH3), 3.30
(3H, s, CH3), 1.83 (3H, s, CH3), 1.70 (3H, s, CH3), 1.51 (6H, s,
CH3). δC (100 MHz, CDCl3): 180.3, 156.4, 154.6, 151.0, 132.2,
130.4, 130.2, 127.3, 122.2, 121.5, 119.3, 116.2, 115.4, 112.6,

4.5.3. Characterization of 4-(5-hydroxy-8,8-dimethyl-10-(3-
methylbut-2-en-1-yl)-4-oxo-4,8-dihydropyrano[3,2-g]chromen-3-
yl)phenyl 4-methylbenzenesulfonate (10)
Yield: 75%. δH (400 MHz, CDCl3): 7.77 (1H, s, C=CH), 7.31
(2H, d, J 8.5 Hz, Ar-H), 6.88 (2H, d, J 8.5 Hz, Ar-H), 6.75 (1H, d,
J 9.9 Hz, C=CH), 5.56 (1H, d, J 9.9 Hz, C=CH), 2.67 (2H, m,
CH2), 1.83 (2H, m, CH2), 1.50 (6H, s, CH3), 1.46 (6H, s, CH3).
δC (100 MHz, CDCl3): 158.2, 151.1, 150.8, 150.1, 131.6, 130.0,
129.4, 129.1, 125.2, 115.8, 115.2, 114.8, 114.1, 105.1, 103.1,
79.8, 33.1, 29.8, 25.1, 17.1. HMRS (ESI) found 405.1669, calc.
405.1624. Anal. found C 74.1, H 6.0. Calc. for C25H24O5: C 74.2,
H 6.0%.
Yield: 58%. δH (400 MHz, CDCl3): 13.01 (1H, s, OH), 7.89
(1H, s, C=CH), 7.78 (2H, d, J 8.4 Hz, Ar-H), 7.49 (2H, d, J 8.7
Hz, Ar-H), 7.36 (2H, d, J 8.4 Hz, Ar-H), 7.10 (2H, d, J, 8.7 Hz,
Ar-H), 6.71 (1H, d, J 9.9 Hz, C=CH), 5.63 (1H, d, J 9.9 Hz,
C=CH), 5.25 (1H, m, C=CH), 3.38 (2H, m, CH2), 2.49 (3H, s,
CH3), 1.83 (3H, s, CH3), 1.69 (3H, s, CH3), 1.50 (6H, s, CH3). δC
(100 MHz, CDCl3): 178.9, 132.5, 131.6, 130.1, 129.8, 128.5,
127.3, 122.4, 122.5, 77.9, 17.9. HMRS (ESI) found 559.2178,
calc. 559.1712. Anal. found C 68.7, H 5.4, S 5.9. Calc. for
C32H30O7S: C 68.8, H 5.4, S 5.7%.
4.6.2. Characterization of 7-(3,4-dihydroxyphenyl)-2,2,10,10-
tetramethyl-11,12-dihydro-2H-dipyrano[2,3-f:2',3'-h]chromen-
8(10H)-one (16, isopomiferin)
Yield: 65%. δH (400 MHz, CDCl3): 7.83 (1H, s, C=CH), 7.24
(1H, d, J 1.9 Hz, Ar-H), 6.88 (1H, d, J 8.3 Hz, Ar-H), 6.77 (1H, d,
J 8.3 Hz, Ar-H), 6.74 (1H, d, J 10.0 Hz, C=CH), 5.58 (1H, d, J
10.0 Hz, C=CH), 2.69 (2H, m, CH2), 1.85 (2H, m, CH2) 1.85 (6H,
s, CH3), 1.47 (6H, s, CH3). δC (100 MHz, CDCl3): 180.7, 160.1,
157.5, 155.6, 153.1, 150.9, 143.2, 141.2, 141.0, 129.2, 118.7,
110.1, 110.0, 108.6, 106.2, 104.3, 100.3, 87.2, 82.4, 33.2, 28.2,
26.5, 16.3. HMRS (ESI) found 421.1653, calc. 421.1573. Anal.
found C 71.4, H 5.8. Calc. for C25H24O6: C 71.4, H 5.8%.
4.5.4 Characterization of 4-(5-hydroxy-8,8-dimethyl-10-(3-
methylbut-2-en-1-yl)-4-oxo-4,8-dihydropyrano[3,2-g]chromen-3-
yl)-1,2-phenylene bis(4-methylbenzenesulfonate) (11)
Yield: 69%. δH (400 MHz, CDCl3): 12.93 (1H, s, OH), 7.92
(1H, s, C=CH), 7.69 (4H, m, J 1.8, Ar-H), 7.48 (1H, m, Ar-H),
7.46 (1H, m, Ar-H), 7.35 (1H, m, Ar-H), 7.33 (4H, m, Ar-H),
6.69 (1H, d, J 9.9 Hz, C=CH), 5.61 (1H, d, J 9.9 Hz, C=CH),
5.25 (1H, m, C=CH), 3,36 (2H, m, CH2), 2.48 (6H, s, CH3), 1.83
(3H, s, CH3), 1.70 (3H, s, CH3), 1.50 (6H, s, CH3). δC (100 MHz,
CDCl3): 159.2, 157.5, 145.7, 141, 132.1, 131.7, 129.8, 128.6,
124.3, 121.7, 77.8, 28.7, 28.1, 25.8, 21.8, 17.9. HMRS (ESI)
found 729.2207, calc. 729.1750. Anal. found C 64.2, H 5.1, S 8.8.
Calc. for C39H36O10S2: C 64.3, H 5.0, S 8.8.
4.7. Biological screening studies
4.7.1. Cytotoxicity drug assay
All test compounds were dissolved in DMSO at a final
concentration of 50 mM and stored at -20°C before use.
Cytotoxicity was assessed by using the 3-(4,5-dimethylthiazole-
2yl)-2,5-diphenyltetrazolium bromide (MTT) (5 mg/ml) assay as
previously described.23 Briefly, 4x103 cells per well were seeded
in 96-well plates before drug treatments. After overnight cell
culture, the cells were then exposed to different concentration of
selected compounds (0.19-100 μM) for 72 hours. Cells without
drug treatment were used as control. Subsequently, 10 μl of 5
mg/mL MTT solution was added to each well and incubated at
37°C for 4 hours followed by addition of 100 μL solubilization
buffer (10 mM HCl in solution of 10% of SDS) and overnight
incubation. A570 nm was then determined in each well on the next
day. The percentage of cell viability was calculated using the
following formula: Cell viability (%) = Atreated/ Acontrol x 100. A
representative graph of at least three independent experiments is
shown in supporting information Table 2-8.
4.5.5. Characterization of 4-(2,2,10,10-tetramethyl-8-oxo-
8,10,11,12-tetrahydro-2H-dipyrano[2,3-f:2',3'-h]chromen-7-
yl)phenyl 4-methylbenzenesulfonate (15)
Yield: 59%. δH (400 MHz, CDCl3): 7.70 (1H, s, C=CH), 7.53
(2H, d, J 8.5 Hz, Ar-H), 7.25 (2H, d, J 8.5 Hz, Ar-H), 6.68 (1H, d,
J 9.9 C=CH), 5.50 (1H, d, J 9.9 Hz, C=CH), 3.07 (3H, s, CH3)
2,61 (2H, m, CH2), 1.75 (2H, m, CH2), 1.42 (6H, s, CH3), 1.45
(6H, s, CH3). δC (100 MHz, CDCl3): 179.6, 159.8, 153.4, 151.9,
150.1, 143.2, 139.2, 130.1, 128.6, 121.4, 117.4, 113.5, 109.8,
102.7, 101.6, 82.9, 82.0, 40.1, 31.3, 29.1, 28.8, 26.5, 24.9, 19.1.
HMRS (ESI) found 559.1717, calc. 559.1720. Anal. found C
68.5, H 5.3, S 5.8. Calc. for C32H30O7S: C 68.4, H 5.2, S 5.8%.
4.6. Synthesis of Isoosajin (14) and isopomiferin (16)
4.7.2 BACE-1 inhibition assay
Isoosajin (14) and isopomiferin (16) were synthesized by
adapting the procedure reported by Wolfrom et al13.The reactions
were carried out on the opportune intermediates, which were
previously isolated from M. pomifera fruits according to a
previously reported procedure9. To perform the cyclization
reaction, a round bottom flask was charged with osajin or
pomiferin and 7 mL of glacial acetic acid. The mixture was
heated to 100 °C and, subsequently, 0.2 mL of sulfuric acid were
added dropwise. The mixture was stirred at room temperature
and the reaction was monitored by TLC (hexane/ethyl acetate
1:1). The reaction was quenched by pouring the mixture in cold
water. The resulting solid was collected by filtration to give the
desired product.
Assay was performed in a 96-well flat bottom white plate
-Secretase Activity Fluorometric Assay Kit (Biovision,
CA, USA). To begin, 5 x 106 cells were collected by
centrifugation with ice-cold extraction buffer added for
homogenization. After centrifugation, 50 µl supernatant (cell
lysate) was transferred to each well in the 96-well plate. 2 µl of
active β-
the 50 µl of extraction buffer as the positive control. For negative
control, 2 µl of the β-secretase inhibitor was added to the 50 µl
sample well. 2 µl of each of the tested samples 4-16 (drug
concentration: IC50 value in BEAS-2B cells-table 1) were added
into each sample well for inhibitory activity evaluation.
Following compounds addition, 50 µl of 2X reaction buffer was
added with gentle mix and incubation for 20 min at 37°C before
adding of 2 µl of β-secretase substrate. The plate was then
covered and incubated in the dark at 37°C for 1 hour. Samples
were then read in a fluorescent 96-well plate reader (Ex/ Em =
4.6.1. Characterization of 7-(4-hydroxyphenyl)-2,2,10,10-
tetramethyl-11,12-dihydro-2H-dipyrano[2,3-f:2',3'-h]chromen-
8(10H)-one (14, isoosajin)

345/500 nm). Background readings produced from substrate
without the addition of secretase were subtracted from all
samples. Secretase activity was calculated as the fold change in
secretase activity. A representative graph of at least three
independent experiments is shown in Fig. 5.
avoid bias due by an unusual or unfavorable orientation of the
aminoacidic residues within the possible interaction site of our
compounds (PDB IDs: 5CLM, 2ZHR, 1M4H), The PDB
structures were optimized for the docking studies. All the
hydrogens were added considering a pH of 5,5 and then the
catalytic dyad composed by Asp32 and Asp228 was controlled
and organized in its physiological condition. After the
preparation of the PDB files of BACE-1 and of the small
molecules, a rough massive screening has been performed with
all the synthetized compounds using the software PyRx
(https://pyrx.sourceforge.io/). In this first part of the virtual
screening, all the three crystal structures of BACE-1 have been
used. The whole surface area of the enzyme was used for the
docking process (blind docking). An average of the energies of
interactions (Kcal/mol) obtained with the three different PDB
structures have been calculated and reported in figure 3b. In the
second round, the 5 compounds with the best docking score and
all the molecules active in the biological tests have been docked
using Autodock Vina25 reducing the research space within a
smaller area around the active site of the enzyme in all the three
different crystal structures. As a method validation, we used the
oxazine molecule co-crystalized in the 5CLM PDB file observing
the pose obtained for this compound and comparing it with that
one of the original crystal structure. With all the adopted
protocols this molecule assumes the same orientation as in the
original crystal structure. The complexes were analyzed using
BIOVIA Discovery Studio26 and the relative 2D interaction plots
were obtained. A similar approach has been adopted for
performing docking studies of our compounds and P-gp. The
molecules were newly minimized at ph 7.2 because the efflux
pump work at physiological pH. The chosen crystal structure is
the 3G60. We used as a positive control Verapamil. (see the
Supporting information for further information about docking
studies).
4.7.3 Rhodamine 123 exclusion assay (P-gp assay)
P-gp activity was determined by measuring intracellular
accumulation of rhodamine 123 in A549 cell lines in the absence
or presence of P-gp inhibitors. Briefly, cells were incubated at
37C with 10µM verapamil, 0.5µM of 1, 0.5µM of 3 or 0.5µM of
6
respectively for
4
hours. Subsequently, 5µg/mL of
rhodamine123 was added to each well and the wells were
incubated for another 1 hour. After washing in phosphate –
buffered saline, cells were lysed in distilled water, and
intracellular levels of rhodamine 123 were quantified by
spectrofluorimetry using CellQuest (BD Biosciences, San Jose,
CA, USA). Data were expressed as % of rhodamine 123
accumulation in control cells not exposed to P-gp inhibitors,
arbitrarily set at 100%. Experimental data were routinely
expressed as means ± SEM from at least three independent
experiments using the Prism software (GraphPad software, La
Jolla, CA, USA).
4.7.4. P-gp Glo activity assay (ATPase assay)
The activity of P-gp ATPase in response to 7 was determined
by Pgp-Glo assay system (Promega, Madison, WI). According to
the manufacture instruction, the stimulatory effect of 7 on the
activity of P-gp ATPase was measured. The luminescence of the
sample reflects the ATP level in the sample, which is positively
correlated with the activity of P-gp ATPase and was recorded
using the SpectraMax Paradigm Multi-Mode Microplate Reader
(Molecular Devices).
Acknowledgements
4.8 Molecular properties and drug-likeness of reference
compounds
This work was supported by a FDCT grant from the Macao
Science and Technology Development Fund to VKWW (Project
code: 0022/2018/A1) and by University of Padova (GZ).
Blood–brain barrier permeation (BBB) was predicted using
online BBB permeation predictor software available at
http://www.cbligand.org/BBB/. As per the software, any
compound having a SVM_MACCSFP BBB score of higher than
0.02 will be considered as able to cross BBB. (Supp. information
Table 1)
Conflict of interest
The authors have declared no conflict of interest.
Appendix A. Supplementary material
4.9 Molecular docking studies
Supplementary data to this article can be found online at https://
All the compounds used for these computational studies have
been drawn with ChemDraw and then minimized with
Avogadro24. For the process of minimization, the appropriate
hydrogen atoms were added to all the drawn compounds,
considering an environmental pH of 5.5, which is the actual pH
at which the enzyme BACE-1 operates. All the structures have
been minimized using the MMFF94 forcefield. For what
concerns the choice of the enzymatic crystal structure used for
the docking assays, several parameters were considered. Firstly,
the crystal structures should be complete in the region near the
active site and in the surrounding areas. Secondly, we have
considered the limitations that this computational method brings
with it: the classical docking approach maintains the
macromolecular structure of the protein rigid and not able to
adapt to the different small molecules that are tested as ligands.
For these reasons, several PDB files have been selected, just to
References:
[1]
[2]
a) M. Wortmann. Alzheimer’s Res. Ther. 2012, 4, 40−43. (b) Alzheimer’s
Disease International. World Alzheimer Report 2015.
http://www.alz.co.uk/research/WorldAlzheimerReport2015.pdf. c) J. Cummings,
G. Lee, T. Mortsdorf, A. Ritter, K. Zhong. Alzheimer Dement (N Y).3 (3), 367-
384. d) https://www.alz.org/national/documents/topicsheet_treatements.pdf
a) K. Beyreuther, C. L. Masters. Brain Pathol. 1991, 1, 241–251. b) J. Hardy , D.
Allsop . Trends Pharmacol Sci. 1991, 12, 383–388. c) D. J Selkoe . Neuron. 1991,
6, 487–498. d) J. A. Hardy, G. A. Higgins. Science. 1992;256:184–185. e) D. J.
Selkoe, J. Hardy. EMBO Mol Med. 2016, 8, 595–608. f) R. Vassar. Alzheimer’s
Res Ther. 2014, 6, 89. g) R. Vassar, P. H. Kuhn, C. Haass, M. E. Kennedy, L.
Rajendran, P. C. Wong, S. F. Lichtenthaler, J Neurochem. 2014, 130, 4–28.
a) S. Filser, S. V. Ovsepian, M. Masana, L. Brazquez-Lorca, A. Brandt Elvanq, C.
Volbracht, M. B. Muller, C. K. Jung, J. Herms. Biol Psychiatry. 2015, 77, 729–
[3]

739. b) K. Zhu, X. Xiang, S. Filser, P. Marinkovic, M. M. Dorokstar, S. Crux, U.
Neumann, D. R. Shimshek, G. Rammes, C. Haass, S. F. Lichtenthaler, J. M.
Gunnersen, J. Herms.. Biol Psychiatry. 2018, 83, 428–437
[14] Y. Hamada, H. Ohta, N. Miyamoto, R. Yamaguvhi, A. Yamani, K.Hidaka, T.
Kimura, K. Saito, Y. Hayashi, S. Ishiurac, Y. Kisoa, Bioorg. Med. Chem. Lett.,
2008, 18 (5), 1643-1647.
[4]
a) Merck, S. (2018). Merck Announces Discontinuation of APECS Study
Evaluating Verubecestat (MK-8931) for the Treatment of People with Prodromal
Alzheimer’s Disease. Available online at: http://investors.merck.com/news/
press-release-details/2018/Merck-Announces-Discontinuation-of-APECSStudy-
Evaluating-Verubecestat-MK-8931-for-the-Treatment-of-Peoplewith-Prodromal-
Alzheimers-Disease/default.aspx (Accessed February 25, 2018). b) E. Malone..
Scrip. 2018 June;12. c) Jansenn. Update on Janssen’s BACE inhibitor program.
Titusville, United States: Company press release; 2018 May 17. d)
https;//www.eisai.com/news/2018/pdf/enews201886pdf.pdf (Accessed October
18, 2018) e) U. Neumann, M. Ufer, L. H. Jacobson, M. L. Rouzade-Dominguez,
G. Huledal, C. Kolly, R. M. Lüönd, R. Machauer, S. J. Veenstra, K. Hurth, H.
Rueeger, M. Tintelnot-Blomley, M. Staufenbiel, D. R. Shimshek, L. Perrot, W.
Frieauff, V. Dubost, H. Schiller, B. Vogg, K. Beltz, A. Avrameas , S. Kretz, N.
Pezous, J. M. Rondeau, N. Beckmann, A. Hartmann, S. Vormfelde, O. J. David,
B. Galli, R. Ramos, A. Graf, C. Lopez Lopez. EMBO Mol Med. 2018, pii: e9316
(November 2018). f) C. Lopez Lopez , A. Caputo, F. Liu, M. E. Riviere, M. L.
Rouzade-Dominguez, R. G. Thomas, J. B. Langbaum, R. Lenz, E. M. Reiman, A.
Graf, P. N. Tariot. J. Prev Alzheimers Dis. 2017, 4, 242–246. g) http://www.ctad-
alzheimer.com/files/files/CTAD%2011th%20Final%20Program%20-
%2017%20oct%20BD-1.pdf (OC-3, 11th Clinical Trials on Alzheimer’s Disease
Conference (CTAD), October 25, 2018).
[15] L. A. Mohammed., J. N. Keller, A. Kaddoumi. Biochim. Biophys Acta. 2016.
1862(4), 778-787.
[16] M. Contino, M. Cantore, E. Capparelli, M. G. Perrone, M. Niso, C. Inglese, F.
Berardi, M. Leopoldo. R. Perrone, N. A. Colabufo. ChemMedChem. 2012. 7,
391-395.
[17] V. K. Wong, Y. K. Law, T.Efferth, O. Kadioglu, L. Liu. AU Patent 101721, 2015.
[18] C. Martin, G. Berridge, C. F. Higgins, R. Callaghan. Br. J. Pharmacol., 1997,
122, 765-771.
[19] N. Colabufo, F. Berardi, M .G. Perrone, E. Capparelli, M. Cantore, C.Inglese,
R.Perrone.Curr. Top. Med. Chem. 2010, 10, 1703-1714.
[20] H. Pajouhesh, G. R. Lenz, NeuroRx. 2005, 2(4): 541–553.
[21] a) T. Suomeni, T. P. Piepponen, R. Kostianein. Plos one, 2015, 10 (7), e0133904;
b) H. Asaba, K. Hosoya, H. Takanaga, S. Ohtsuki, E. Tamura, T. Takizawa. J.
Neurochem, 2000, 75, 1907-1916; c) M.Miyajima, H. Kusuhara, M. Fujishima, Y.
Adachi, Y. Sugiyama, Drug. Metab. Dispos, 2011, 814-819.
[22] a) F. J. Rombouts, G. Tresadern, O. Delgado, C. Martinez-Lamenca, M. Van
Gool, A. Garcia-Molina, S. A. Alonso de Diego, D. Oehlrich, H. Prokopcova, J.
M. Alonso, N. Austin, H. Borghys, S. Van Brandt, M. Surkyn, M. De Cleyn, A.
Vos, R. Alexander, G. Macdonald, D. Moechars, H. Gijsen, A. A. Trabanco, J.M.
Chem, 2015, 58, 8216-8235. b) N. Maki, P. Hafkemeyer, S. Dey. J. Biol. Chem.
2003, 16, 278 (20), 18132-9.
[5]
D. Kuhnke, G. Jedlitschky, M. Grube, M. Krohn, M. Jucker, I. Mosyagin, I.
Cascorbi, L. C. Walker, H. K. Kroemer, R. W. Warzok, S. Vogelgesang. Brain
Pathol. 17(4), 2011, 347-353
[23] V. K. Wong, T. Li, B. Y. Law, E. D. Ma, N. C. Yip, F. Michelangeli, C. K. Law,
M. M. Zhang, K. Y. Lam, P. L. Chan, L. Liu, Cell Death Dis. 2013, 4, e720.
[24] M. D. Hanwell, D. E. Curtis, D. C. Lonie, T. Vandermeersch, E. Zurek and G. R.
Hutchison, Journal of Cheminformatics, 2012, 4:17.
[6]
[7]
E. Zanforlin, G. Zagotto , G.Ribaudo, Current Medicinal Chemistry, 2017,
24(34), 3749-3773.
[25] O. Trott, A.J. Olson, J. Comput. Chem., 2010, 31, 455–461.
[26] BIOvIA, D. S. (2015). Discovery studio modeling environment. San Diego,
Dassault Systemes, Release, 4.
a) S. S. Lim, S. Han, S. Y. Kim, Y. E Bae, I. J. Kang. Food Sci Biotech. 2007, 16,
265 b) M. Jung, M. Park, Molecules. 2007. 12, 2130. c) H. Kim, B.S Park, K. G.
Lee, C.Y.Choi, S.S.Yang, Y.H.Kim, S.E.Lee, J.Agric.Food Chem. 2005, 53,
8537. d) J. Zhu, R. Choi, G.Chu, A. Cheung, Q. Gao, J. Li, Z.Y. Jiang, T. Dong,
K. Tsim, J. Agric.Food Chem. 2007, 55, 2438.
[27] Y. S. Ki, E. Y. Park, H. – W. Lee, M. S. Oh, Y.- W. Cho, Y. K. Kwon, J. H.
Moon, K. – T. Lee. Biol. Pharm. Bull. 2010, 33 (6), 1054-1058.
[8]
[9]
a) K. Youn, J-H. Park, J. Lee, W-S. Jeong, C-T. Ho, M. Jun. Nutrients. 2016.8,
637; b) K Youn, J-H. Park, S. Lee, S Lee, J. Lee, E-Y Yun, W-S. Jeong, M. Jun.
2018. J.Med. Food. 00 (0), 1-5.
G Ribaudo, M. A. Pagano, V. Pavan, M. Redaelli, M. Zorzan, R. Pezzani, C.
Mucignat-Caretta, T. Vendrame, S. Bova, G. Zagotto, Fitoterapia. 105:132–138
Supplementary Material
[10] a) D. Zhao, C. Zhao, X. Chen, H. Xia, L.Zhang, H. Liu, X. Jiang, Y.Dai, J. Liu.
Neurochem Res.2013. 2105-2113. b) N. H. Mohamed, H. M. Ali, M. Z. M.
Salem. J Pure Appl. Microbiol. 2014, 8, 2969- 2974. c) H. I. Moon. Nat. Prod.
Commun. 2014, 9, 1723-1724
Supplementary material that may be helpful in the review
process should be prepared and provided as a separate electronic
file. That file can then be transformed into PDF format and
submitted along with the manuscript and graphic files to the
appropriate editorial office.
[11] E. Kupeli, I. Orhan, G.Toker, E. Yesilada, J Ethnopharmacol. 107:169–174.
[12] a) G. Delle Monache, R. Scurria, A. Vitali, B. Botta, B. Monacelli, G. Pasqua, C.
Palocci, E. Cernia., Phytochemistry. 1994, 37:893–898. b) G. Ribaudo, T.
Vendrame, S. Bova, Nat. Prod. Res. 2017, 1988-1994.
[13] M. L. Wolfrom, F. L. Benton, A. S. Gregory, W. W. Hess, J. E. Mahan, P. W.
Morgan, J. Am. Chem. Soc., 1939, 61 (10), pp 2832–2836.
Click here to remove instruction text...

Highlights:

Twelve semi-synthetic derivatives of
prenylated isoflavones 5-16 were designed
and synthesized.

Derivatives were evaluated via human
recombinant BACE-1 inhibition and
cytotoxicity assay.


7, 8 and 13 were found to be the most
active candidates.
These compounds were targeted for the
BACE-1 enzyme and P-glycoprotein (P-gp)
in docking studies.

7 enhance the activity of P-glycoprotein (P-
gp) on A549 cancer cells.