Multitarget drug design strategy against Alzheimer's disease: Homoisoflavonoid Mannich base derivatives serve as acetylcholinesterase and monoamine oxidase B dual inhibitors with multifunctional properties

Article abstract of DOI:10.1016/j.bmc.2016.11.048

A series of homoisoflavonoid Mannich base derivatives were designed, synthesized and evaluated as multifunctional agents against Alzheimer's disease. It demonstrated that most of the derivatives were selective AChE and MAO-B dual inhibitors with good mult

Full text of DOI:10.1016/j.bmc.2016.11.048

Accepted Manuscript
Multitarget drug design strategy against Alzheimer’s disease: Homoisoflavo-
noid Mannich base derivatives serve as acetylcholinesterase and monoamine
oxidase B dual inhibitors with multifunctional properties
Yan Li, Xiaoming Qiang, Li Luo, Xia Yang, Ganyuan Xiao, Yunxiaozhu Zheng,
Zhongcheng Cao, Zhipei Sang, Fu Su, Yong Deng
PII:
S0968-0896(16)30729-5
DOI:
http://dx.doi.org/10.1016/j.bmc.2016.11.048
BMC 13415
Reference:
Bioorganic & Medicinal Chemistry
To appear in:
Received Date:
Revised Date:
Accepted Date:
16 September 2016
22 November 2016
25 November 2016
Please cite this article as: Li, Y., Qiang, X., Luo, L., Yang, X., Xiao, G., Zheng, Y., Cao, Z., Sang, Z., Su, F., Deng,
Y., Multitarget drug design strategy against Alzheimer’s disease: Homoisoflavonoid Mannich base derivatives serve
as acetylcholinesterase and monoamine oxidase B dual inhibitors with multifunctional properties, Bioorganic &
Medicinal Chemistry (2016), doi: http://dx.doi.org/10.1016/j.bmc.2016.11.048
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.

Multitarget drug design strategy against Alzheimer’s disease: homoisoflavonoid
Mannich base derivatives serve as acetylcholinesterase and monoamine oxidase B
dual inhibitors with multifunctional properties
Yan Lia, Xiaoming Qianga, Li Luoa, Xia Yanga, Ganyuan Xiaoa, Yunxiaozhu Zhenga, Zhongcheng
Caoa, Zhipei Sangb, Fu Sua,*, Yong Denga,*
aDepartment of Medicinal Chemistry, Key Laboratory of Drug Targeting and Drug Delivery System
of the Education Ministry, West China School of Pharmacy, Sichuan University, Chengdu, 610041,
China
bCollege of Chemistry and Pharmaceutical Engineering, Nanyang Normal University, Nanyang,
473061, China
*Corresponding Author.
Phone, +86-28-85501310; E-mail: sufu@scu.edu.cn (Fu Su)
Phone, +86-28-85503790; E-mail: dengyong@scu.edu.cn (Yong Deng)
Abstract

A series of homoisoflavonoid Mannich base derivatives were designed, synthesized and evaluated as
multifunctional agents against Alzheimer’s disease. It demonstrated that most of the derivatives were
selective AChE and MAO-B dual inhibitors with good multifunctional properties. Among them,
compound 10d displayed the comprehensive advantages, with excellent AChE and MAO-B
inhibitory activities (IC50 = 2.49 ± 0.08 nM and 1.74 ± 0.0581 μM, respectively), good self- and
Cu2+-induced Aβ1-42 aggregation inhibitory potency, antioxidant activity, biometal chelating ability
and high BBB permeability. These multifunctional properties make 10d as an excellent candidate for
the development of efficient drugs against AD.
Keywords:
Alzheimer’s disease; homoisoflavonoid; Mannich base derivatives; acetylcholinesterase inhibitors;
monoamine oxidase B inhibitors; multifunctional properties.
1. Introduction

Alzheimer’s disease (AD) is a progressive and irreversible neurodegenerative disorder that
characterized by memory loss, behavioral disturbances and many other cognitive impairments. As
reported by WHO, more than 30 million people are afflicted with AD worldwide, and this number is
expected to increase to 100 million by 2050.1 Unfortunately, the exact etiology remains unknown and
the drugs approved for AD clinical treatment only target the symptoms, but cannot mitigate the
progression of AD. It is acknowledged that AD possesses an extremely complex interconnected
network of multiple factors, including β-amyloid (Aβ) deposits, τ-protein aggregation, oxidative
stress, dyshomeostasis of biometals and degeneration of cholinergic neurons in the central nervous
system (CNS).2
According to the classical “cholinergic hypothesis”, the impairment of the cholinergic
neurotransmission in brain, resulting from a substantial acetylcholine (ACh) deficiency, leads to the
cognitive impairments in AD patients. This hypothesis generally implies that the decreased ACh
metabolism will improve the cognition in AD.1 ACh is mainly hydrolyzed by acetylcholinesterase
(AChE, EC 3.1.1.7) at the cholinergic synapses. Therefore, several AChE inhibitors such as
donepezil, rivastigmine, galantamine and Huperzine A have been approved by FDA for AD treatment.
These drugs produce clinically relevant benefits in all clinical domains of AD. And AChE has always
been a popular target in anti-AD research. Moreover, AChE is also implicated in AD pathogenesis
through influencing the Aβ deposition. Studies pointed out that AChE could act as a molecular
chaperone, accelerating the formation and aggregation of Aβ fibrils and further increasing their
neurotoxicity, through the peripheral anionic site (PAS).3 In addition, it is supposed that
butyrylcholinesterase (BuChE, EC 3.1.1.8) may have a compensatory role in the modulation of ACh
hydrolysis. However, BuChE is primarily localized in plasma, liver and muscle tissues, and the
pharmacological role of which was not yet completely understood.4 Accordingly, serious inhibition
of BuChE may contribute to the potential peripheral side effects of ChE inhibitors. Therefore,
developing of novel AChE inhibitors, with expectation of lower side effects and certain selectivity,
may be more rational for anti-AD drug searching.
Generally, the clinicopathological features of AD are the senile plaques (SPs) and
neurofibrillary tangles (NFTs) in the brain. The major components of the SPs are Aβ peptides, which
exist as Aβ1-40 and Aβ1-42. According to the “amyloid cascade hypothesis”, these peptides can form
the neurotoxic aggregates and then trigger the pathogenic cascade, and finally cause the neuronal loss

and dementia.5 Among the two isoforms, Aβ1-42 displays lower solubility, more neurotoxicity and
higher tendency to aggregate.6 Therefore, preventing the formation and aggregation of Aβ1-42 could
provide neuroprotective effects and a potential efficient strategy against AD.
Additionally, remarkably high concentrations of biometals, such as Cu2+, Fe2+, Zn2+ and Al3+,
have been found in the amyloid deposits in AD-affected brains. Several studies have indicated that
these metal ions could promote the aggregation of Aβ as well lead to the formation of reactive
oxygen species (ROS) and oxidative stress.7 While, abundant human data consistently support that
oxidative stress occurs as a constant feature in AD brain pathology.8 The “oxidative stress hypothesis”
states that the increased formation of ROS and free radicals leads to the deleterious effects on cell
components and may likely end in severe mitochondria damage and cellular death.9 Furthermore,
studies showed that Aβ neurotoxicity mainly depends on the catalytic generation of H2O2, which
could produce the oxidative stress. In a vicious cycle, the oxidative stress could in turn promote the
Aβ aggregation by generating modified Aβ species which are prone to aggregate and resistant to
elimination.10 Collectively, the biometals homeostasis and efficient antioxidant protection are of
great importance in AD patients.
Recently, monoamine oxidases (MAOs), the important factors involved in oxidative stress, have
attracted more and more attention. MAOs are flavin adenine dinucleotide (FAD)-containing enzymes,
which exist as MAO-A and –B. They are responsible for the regulation and metabolism of major
monoamine neurotransmitters and xenobiotic amine. However, the biochemical activity of MAOs
generates hydroxyl radicals, the toxic members of the oxygen free radical group.11 So it seems
beneficial for AD treatment to inhibit the enzymes activities. While, studies suggested that MAO-A
inhibitors may precipitate a potentially fatal hypertensive response when combined with foods
containing tyramine and this may be the major cause of the restricted therapeutic use of MAO-A
inhibitors.12,13 It is established that the activity of MAO-B could increase up to 3-fold in the temporal,
parietal, and frontal cortex of AD patients compared with controls.14 And evidence shows that as
aging progresses MAO-B activity increases in brain, but not so for MAO-A.15 On the other hand,
MAO-B inhibitors such as selegiline and rasagiline have shown the abilities to improve learning and
memory deficits in AD animal models and to slow down the progression of AD in patients.14
Therefore, MAO-B inhibitory activity should be an important factor to consider during rational
anti-AD drug design.

Due to the pathological complexity of AD, the current therapeutic drugs and the traditional
“one-molecule, one-target” paradigm seem not so efficient. According to the popular
“multi-target-directed ligands” (MTDLs) strategy, it is advantageous to develop drugs with two or
more complementary bioactivities for the treatment of complex diseases as AD. Considering above,
our work is focused on the rational design, synthesis and evaluation of novel multifunctional
molecules against AD.
Homoisoflavonoids, 3-benzylidene-4-chromanones, are naturally occurring products that related
to flavonoids and exhibit various biological activities (Figure 1a). Many studies have indicated that
homoisoflavonoids and the synthesized analogs showed high affinities toward Aβ aggregates16 and
good selective MAO-B inhibitory activities17,18. Furthermore, homoisoflavonoids have been reported
to possess antioxidant19,20 and anti-inflammatory21 properties. The diversity of the bioactivity makes
homoisoflavonoid an excellent leading scaffold to design novel multifunctional agents against AD. In
addition, molecules bearing phenolic Mannich base moieties (Figure 1b) have displayed good
antioxidant22,23, anti-inflammatory24 and AChE inhibitory25,26 activities. These results provide
valuable suggestions for the anti-AD drug design. As mentioned above, AChE inhibitors produce
clinically relevant benefits in all clinical domains of AD. Considering the important role of AChE in
the AD pathology and the clinical efficacy of AChE inhibitors, the pharmacophores of donepezil
(Figure 1c) were taken as references during the drug design. Collectively, we designed and
synthesized a series of homoisoflavonoid Mannich base derivatives that are expected to act as
multifunctional agents with AChE inhibition, anti-Aβ aggregation, MAO-B inhibition, antioxidant
and biometal chelating properties (Figure 1d).

Figure 1. Design strategy of homoisoflavonoid Mannich base derivatives.
2. Results and discussion
2.1. Chemistry
The synthetic pathways of target derivatives were summarized in Scheme 1. Firstly, Mannich
reactions of compounds 1 and 2 with paraformaldehyde and different secondary amines in EtOH
gave the Mannich base intermediates 3a-k, 4a-e and 5. The different reactivity allows the
regioselectivity for Mannich reaction. The exclusive formation of products with aminoalkylation
adjacent to the hydroxyl group is attributed to the directing effect and the hydrogen bonding between
the nitrogen and the o-hydroxyl group, that afford a stabilized six-membered ring27. The key
intermediate 6,7-dimethoxychroman-4-one (8), was obtained from 3,4-dimethoxyphenol (6)28 after
Fries rearrangement and the subsequent cyclization. Condensations of 8 with appropriate substituted
benzaldehydes 1, 3a-k and 4a-e in ethanolic 50% KOH solution, respectively, afforded the

corresponding homoisoflavonoids 9, 10a-k and 11a-e. Differently, compound 12 was obtained from
8 and 5 through the 85% H3PO4 catalyzed condensation. The sodium salt formation of 4’-OH in
compound 5 reduced the reactivity of aldehyde group, which might lead to the failure of the 50%
KOH catalyzed condensation. As reported in many literatures, only E stereoisomers were isolated
with both procedures. The stereochemistry of the homoisoflavonoid diastereomers has been
elucidated by NMR spectroscopic measurements. Due to the deshielding effect of the carbonyl group,
1
the vinyl protons should give a H NMR signal at a greater chemical shift than in the Z isomer, the
value is about δ 7.60-7.89. Additionally, the protons of the methylene adjacent to double bond are
also useful in distinguishing the isomers. In E-isomers, the signals appeared in the range of δ
5.08-5.35 and in Z-isomers these protons appeared at about δ 4.93-4.9817,19,29-31. All derivatives were
1
characterized by H NMR and HR-ESI-MS, and most of them were further characterized by 13C
NMR.
Scheme 1. Synthesis of homoisoflavonoid Mannich base derivatives 9-12. Reagents and conditions:
(a) paraformaldehyde, secondary amine, EtOH, reflux for 24-48 h; (b) ClCH2CH2COCl, BF3•Et2O,
60 °C, 5 h; (c) K2CO3, EtOH, r.t., overnight; (d) substituted benzaldehyde 1, 3a-k or 4a-e, 50% KOH,
EtOH, r.t., overnight; or for 12, compound 5, 85% H3PO4, 80 °C, 10 h.

2.2. Pharmacology
2.2.1. Evaluation of AChE and BuChE inhibitory activities
All the synthesized homoisoflavonoid Mannich base derivatives were evaluated for the in vitro
AChE and BuChE inhibitory activities, using the modified Ellman’s method32. Donepezil was used
as the reference compound. And the results are summarized in Table 1.
Most of the derivatives exhibited significant EeAChE inhibitory activities with IC50 values
ranging from nanomolar to micromolar. The structure-activity relationship analysis suggested that
the amine moiety and its substituted position exerted great influence on the potency. Compared the
IC50 values of 10a-k with 9, 11a-e and 12, it demonstrated that the introduction of Mannich base
moiety dramatically improved the EeAChE inhibitory activity of homoisoflavonoid and the
4’-aminoalkylated compounds (10a-k) were more appropriate for the potency. Generally, the
aliphatic amine substituted compounds (10a-f) possessed greater AChE inhibitory activities than the
benzylamine substituted compounds (10g-k). And compound 10d displayed the most potent EeAChE
inhibitory activity (IC50 = 0.00249 ± 0.00008 μM), that was more excellent than donepezil (IC50 =
0.0226 ± 0.001 μM) under our experimental condition.
The BuChE inhibition results showed that all the homoisoflavonoid Mannich base derivatives
are AChE selective inhibitors. Considering the potential role of BuChE in AD pathogenesis, to some
extent, this may be a shortcoming. But, the selectivity profile of our derivatives may contribute to the
reduction of the associated side effects and benefit the AD treatment. Among the derivatives, 10k
exhibited the most potent inhibitory activity for BuChE (44.6 ± 0.3% at 50 μM).
Table 1. In vitro inhibition of AChE and BuChE and Oxygen Radical Absorbance Capacity (ORAC,
Trolox Equivalents) by homoisoflavonoid Mannich base derivatives and donepezil.
Compd.
NR3R4 IC50 EeAChE (μM)a
% Inhibition of RatBuChEb
ORACc
—
A
B
C
D
E
F
G
H
I
59.3 ± 1.19
6.18 ± 0.2
n.a.d
1.73 ± 0.02
1.71 ± 0.03
1.97 ± 0.05
1.65 ± 0.01
2.25 ± 0.07
1.35 ± 0.02
2.11 ± 0.04
1.57 ± 0.06
1.78 ± 0.02
1.13 ± 0.03
9
0.536 ± 0.016
0.00680 ± 0.00028
0.362 ± 0.0112
0.00249 ± 0.00008
1.29 ± 0.0398
6.26 ± 0.122
1.99 ± 0.099
16.4 ± 0.885
0.364 ± 0.0076
10a
10b
10c
10d
10e
10f
10g
10h
10i
7.72 ± 0.6
24.1 ± 0.4
19.8 ± 0.6
n.a.d
n.a.d
n.a.d
11.2 ± 0.1
n.a.d

15.7 ± 0.629
18.1 ± 0.882
37.4 ± 0.821
85.9 ± 1.59
84.1 ± 1.68
33.3 ± 0.855
172 ± 6.22
22.7 ± 0.9
44.6 ± 0.3
6.31 ± 0.2
7.85 ± 0.4
n.a.d
1.80 ± 0.06
0.97 ± 0.01
2.05 ± 0.05
1.52 ± 0.07
1.83 ± 0.02
1.69 ± 0.01
1.34 ± 0.03
2.32 ± 0.06
—
10j
10k
11a
11b
11c
11d
11e
12
J
K
B
E
F
H
I
16.0 ± 0.2
n.a.d
0.928 ± 0.0288
0.0226 ± 0.001
n.a.d
D
—
Donepezil
IC50 = 21.4 ± 1.01 μM
a
From Electrophorus electricus. IC50 values represent the concentration of inhibitor required to
decrease enzyme activity by 50% and are the mean of 3 independent experiments, each performed in
triplicate. Data were expressed as mean ± SD (SD = standard deviation ).
b From rat serum. Tested compounds were used at 50 μM.
c
The mean ± SD of the three independent experiments. The data are expressed as μM of Trolox
equiv/μM of tested compound.
d n.a. = no active. Compounds defined “no active” means that percent inhibition is less than 5.0% at a
concentration of 50 μM in the assay conditions.
2.2.2. Kinetic study for the inhibition of AChE
To further investigate the inhibitory mechanism of our derivatives, a kinetic study of EeAChE
was conducted with 10d, the most potent AChE inhibitor among these derivatives. The graphical
presentation of the reciprocal Lineweaver-Burk plots was displayed in Figure 2. As it shown, with
the concentrations of 10d increasing, the slopes (decreased Vmax) and intercepts (higher Km)
increased. And the intersection point of trendlines fell in the third quadrant. This result demonstrated
a mixed-type inhibitory behavior for 10d, as the consequence of a simultaneous binding to the CAS
and PAS of AChE.

Figure 2. Kinetic study on the mechanism of EeAChE inhibition by compound 10d. Overlaid
Lineweaver-Burk reciprocal plots of AChE initial velocity at increasing substrate concentration
(0.1-0.4 mM) in the absence and presence of 10d are shown. Lines were derived from a weighted
least-squares analysis of data points.
2.2.3. Molecular modeling study of AChE
To explore the binding modes of the homoisoflavonoid Mannich base derivatives with AChE,
docking simulation of compound 10d were performed using the docking program, AutoDock 4.2
package with Discovery Studio 2.5, based on the X-ray crystal structure of Torpedo californica (Tc)
AChE (PDB code: 1EVE)33. As shown in Figure 3, 10d occupied the entire enzymatic CAS, the
mid-gorge sites and the PAS. Three hydrogen bonds were generated between the ligand 10d and the
residues Tyr121, Phe331 and His440, respectively (10d-O1• • •HO-Tyr121; 10d-C=O• •
•HN-Phe331; 10d-OH• • •N-His440). The B ring of 10d was observed to adopt a parallel π-π
interaction with Phe330. Moreover, the piperidine moiety and the C ring potentially induced
hydrophobic interactions with Trp84 and Tyr334 respectively. Other interactions could be observed
between the ligand and Gly117, Gly118, Ser122, Gly123, Tyr130, Ser200, Trp279, Ile287 and
Gly335. The docking study indicated that 10d could stably bind to the CAS and PAS of AChE and
explained the high AChE inhibitory potency of our derivatives.

Figure 3. Docking pose of compound 10d (grey stick) into the binding site of TcAChE (1EVE),
highlighting the protein residues that participate in the main interactions with the inhibitor. Picture
was generated with Discovery Studio 2.5. (For interpretation of the references to color in this figure
legend, the reader is referred to the web version of this article.)
2.2.4. In vitro antioxidant activity assay
The in vitro antioxidant activities of the homoisoflavonoid Mannich base derivatives were
determined by a well-established oxygen radical absorbance capacity assay using fluorescein
(ORAC-FL)34. Trolox, a water-soluble vitamin E analogue, was used as a standard, and the
antioxidant activity was expressed as Trolox equivalent. As shown in Table 1, most of the target
compounds exhibited good antioxidant activities with ORAC-FL values of 0.97-2.32 Trolox
equivalents. Compound 10d and 12 showed the most potent antioxidant activities with ORAC-FL
values of 2.25 ± 0.07 and 2.32 ± 0.06 Trolox equivalents, respectively. However, the substituted
positions of the hydroxyl group and the amines showed no significant influence on the potency.
2.2.5. Inhibition of self- and Cu2+-induced Aβ1-42 aggregation

The abilities of the target compounds to inhibit self- and Cu2+-induced Aβ1-42 aggregation were
assessed using the thioflavin T (ThT) fluorescence assay, with curcumin as a reference compound35,36
The results summarized in Table 2 indicated that the introduction of Mannich base moiety improved
both self- and Cu2+-induced Aβ1-42 aggregation inhibitory activities of homoisoflavonoid (9).
Noticeably, the activities of 10b, 10e, 10f, 10h and 10i were higher than those of 11a-e, respectively,
.
which revealed that 3’-OH and 4’-aminoalkylated substitution were more appropriate for the potency.
Different from the AChE inhibition assay, the Aβ1-42 aggregation inhibitory potency of 12 was a little
bit higher than that of 10d. From the results of 10a-k, we could find that, generally, the benzylamine
substituted compounds (10g-k) possessed greater Cu2+-induced Aβ1-42 aggregation inhibitory
activities than the aliphatic amine substituted compounds (10a-f). However, no similar conclusion
was drawn from the self-induced Aβ1-42 aggregation assay. Among the derivatives, 10k was the most
potent inhibitor against both self- and Cu2+-induced Aβ1-42 aggregation (59.6 ± 2.7% and 79.9 ± 2.5%,
respectively), which was more excellent than curcumin (41.3 ± 0.9% and 67.2 ± 1.3% respectively)
under our experimental condition.
Table 2. In vitro inhibition of self- and Cu2+-induced Aβ1-42 aggregation and MAO inhibitory
activities of homoisoflavonoid Mannich base derivatives and reference compounds.
%Inhibition of Aβ1-42
MAO-Ad,f
MAO-Bd,f
aggregationa
Compd.
NR3R4
SIe
Self-inducedb,f Cu2+-inducedc,f
% Inhibition
21.9 ± 1.1
8.2 ± 0.2
9.5 ± 0.2
16.4 ± 0.9
24.6 ± 1.0
23.0 ± 1.2
40.5 ± 2.2
3.4 ± 0.1
12.3 ± 0.4
15.0 ± 0.9
14.0 ± 0.6
28.3 ± 1.1
8.5 ± 0.4
2.1 ± 0.1
6.8 ± 0.3
9.8 ± 1.1
9.8 ± 1.3
IC50 ± SD (μM)
0.19 ± 0.0038
0.882 ± 0.0255
2.43 ± 0.112
0.722 ± 0.0125
1.74 ± 0.0581
1.01 ± 0.0495
0.915 ± 0.0322
1.15 ± 0.0345
1.03 ± 0.0312
1.72 ± 0.0922
1.5 ± 0.0231
1.34 ± 0.0772
36.7 ± 1.0%
9
—
A
B
C
D
E
F
20.1 ± 1.0
35.2 ± 1.2
21.2 ± 0.9
37.3 ± 1.1
39.3 ± 2.1
28.3 ± 1.5
41.9 ± 1.2
7.7 ± 0.2
30.3 ± 1.2
42.2 ± 2.1
40.6 ± 2.5
36.5 ± 1.7
47.0 ± 3.2
38.7 ± 1.9
40.5 ± 2.4
42.6 ± 2.2
67.8 ± 3.2
49.9 ± 1.9
65.8 ± 3.0
79.9 ± 2.5
29.1 ± 0.8
32.4 ± 1.9
36.3 ± 2.2
54.9 ± 3.1
28.0 ± 0.7
> 105.3
> 22.7
> 8.2
> 27.7
> 11.5
> 19.8
> 21.9
> 17.4
> 19.4
> 11.6
> 13.3
> 14.9
—
10a
10b
10c
10d
10e
10f
10g
10h
10i
G
H
I
37.8 ± 1.2
32.7 ± 0.9
44.0 ± 2.1
59.6 ± 2.7
20.1 ± 0.5
7.8 ± 0.4
10j
10k
11a
11b
11c
11d
11e
J
K
B
E
F
18.6 ± 0.9%
—
26.8 ± 1.5
35.6 ± 2.1
8.0 ± 0.9
29.8 ± 1.1%
—
H
I
43.9 ± 2.0%
—
38.1 ± 1.5%
—

12
D
41.4 ± 2.1
49.0
67.2 ± 1.3
—
17.0 ± 0.7
—
0.0027 ± 0.0006 μMg
1.42 ± 0.015 μMg
2.56 ± 0.023 μMg
14 ± 0.232
—
> 1.4
—
Curcumin
Clorgyline
Rasagiline
Iproniazid
—
—
—
—
41.3 ± 0.9
—
—
—
4.19 ± 0.101
0.0825 ± 0.002
1.95 ± 0.074
0.00064
17.21
1.31
—
—
a For inhibition of Aβ aggregation, the thioflavin-T fluorescence method was used.
b Inhibition of self-induced Aβ1-42 aggregation (25 μM) by tested inhibitors at 25 μM.
c Inhibition of Cu2+-induced Aβ1-42 aggregation. The concentration of tested compounds and Cu2+
were 25 μM.
d Percentages are the percent inhibition of MAO by tested compounds at 20 μM.
e SI = IC50 (MAO-A) / IC50 (MAO-B)
f The mean ± SD of the three independent experiments.
g IC50 values of the drugs.
2.2.6. Recombinant human MAO-A and -B inhibition studies
To further study the multipotent biological profile of our derivatives, the MAO-A and -B
inhibitory properties were evaluated with the recombinant human MAO-A and -B as the enzyme
sources. The enzyme activities were determined by a fluorimetric method with kynuramine as a
common substrate37-39. Clorgyline, rasagiline and iproniazid were used as reference drugs. The
derivatives were first tested at the concentration of 20 μM for preliminary screening. Then the
compounds with the inhibition ratios greater than 50% would be retested for the IC50 values at
various concentrations. The results were summarized in Table 2. As it shown, all the target
compounds displayed poor MAO-A inhibitory activities at the concentration of 20 μM. On the
contrary, most of them (except for 11a-e) exhibited remarkable MAO-B inhibitory activities with the
IC50 values ranging from 0.19-14 μM. As mentioned in introduction, this high MAO-B selectivity
profile might be helpful for avoiding the potentially fatal hypertensive response and beneficial for
AD treatment. From the results of 9 and 10a-k, we could find that the introduction of Mannich base
moiety lowered the MAO-B inhibitory potency of homoisoflavonoid. However, this might be
innocuous because of the multifunctional properties obtained in turn. The significant difference in
MAO-B inhibitory activities of 10a-k, 11a-e and 12 revealed that 3’-OH and 4’-aminoalkylated
substitution (10a-k) were more appropriate for the potency, which were consistent with the
conclusions drawn from the AChE and Aβ1-42 aggregation inhibition assays. In addition, the

difference in the substituted amines (NR3R4) showed no obvious impact on the MAO-B inhibitory
potency. Among the Mannich base derivatives, compound 10c was the most potent MAO-B inhibitor
with an IC50 value of 0.722 ± 0.0125 μM.
2.2.7. Molecular modeling studies of MAOs
Similar to the AChE molecular modeling study, docking simulations were performed to explore
the hypothetical binding modes of the homoisoflavonoid Mannich base derivatives with respect to
both isoforms of human MAOs. Taking into account the results of the above assays, 10d was selected
as the optimal representative compound for the docking studies based on the X-ray crystal structures
of human MAO-A (PDB code: 2Z5X) and MAO-B (PDB code: 2V60). The most stable binding
modes of 10d in hMAO-A and –B active sites were graphically inspected (Figure 4). As shown in
Figure 4A, the hMAO-A best pose of the 10d chromanone ring was located far from the enzymatic
cofactor FAD and no hydrogen bond or parallel π-π interaction was observed between the ligand and
the active site, which explained the poor MAO-A inhibitory activities of these derivatives. However,
in MAO-B docking simulation (Figure 4B), the chromanone ring of 10d was close to the FAD
cofactor. Two hydrogen bonds were generated between the ligand 10d and the active site of MAO-B:
one between the 4-carbonyl group and Tyr435 (10d-C=O • • •HO-Tyr435) and the other one between
the 6-OCH3 group and FAD (10d-O • • •HN-FAD). And the A ring of 10d was observed to adopt a
parallel π-π interaction with Tyr398. Meanwhile, other interactions could be observed between 10d
and Pro102, Pro103, Pro104, Trp119, Leu167, Phe168, Leu171, Cys172, Ile198, Ile199, Gln206,
Ile316, Tyr326, Phe343 and so on. Thus, the docking studies explained the high MAO-B inhibitory
activities and selectivities of our derivatives.
Figure 4. Representation of compound 10d docked into the binding sites of hMAO-A (A) and –B

(B), highlighting the protein residues that participate in the main interactions with the inhibitor.
Ligand 10d and FAD are displayed in grey and red respectively. Hydrogen-bonds are shown with the
green dotted lines. Picture was generated with Discovery Studio 2.5. (For interpretation of the
references to color in this figure legend, the reader is referred to the web version of this article.)
2.2.8. Metal-chelating studies
The metal-chelating abilities of the homoisoflavonoid Mannich base derivatives were
investigated by UV-visual spectrometry, with 10d and 10k as representative compounds. The results
were shown in Figure 5. It could be seen that, the spectrum of 10d was significantly changed upon
the addition of CuCl2, FeSO4 and AlCl3 (Figure 5A). The maximum absorption at 302 nm shifted to
300 nm, 300 nm and 298 nm respectively and the absorption increased. And the absorption peak at
378 nm shifted to 380 nm, 372 nm and 380 nm respectively. It revealed a possible interaction
between 10d and these biometals. No significant change was observed upon the addition of ZnCl2.
Similar to 10d, the intensities and positions of the absorption peaks of 10k were changed upon the
addition of CuCl2, FeSO4 and AlCl3, which indicated the possible interaction between 10k and these
biometals (Figure 5B).
Figure 5. UV spectra of compound 10d (A) and 10k (B) (37.5 μM in methanol) alone or in the
presence of CuCl2, FeSO4, ZnCl2 or AlCl3 (37.5 μM, in methanol).
2.2.9. In vitro blood-brain barrier permeation assay
The ability to cross the blood-brain barrier (BBB) and reach the therapeutic targets is an
essential property for anti-AD drugs. The parallel artificial membrane permeation assay of the
blood-brain barrier (PAMPA-BBB), described by Di et al40,41, was conducted to evaluated the BBB
permeability of our derivatives. To validate the assay, the experimental permeabilities of 11

commercial drugs were compared with the reported values. And a plot of experimental data versus
the bibliographic values displayed a good linear correlation: Pe (exp.) = 0.9163 × Pe (bibl.) − 0.2247
(R2 = 0.9558) (Figure 6). With the consideration of the limit established by Di et al.40, we
determined that molecules with Pe values above 3.44 × 10-6 cm/s would be able to cross the BBB
(Table 4)42. The results in Table 5 indicated that compound 10c, 10d, 10j and 10k could penetrate
into brain with good BBB permeabilities.
Table 3. Permeability Pe (× 10-6 cm/s) in the PAMPA-BBB assay for 11 commercial drugs used in
the experiment validation.
Commercial drugs
Verapamil
Oxazepam
Diazepam
Bibla
16
10
16
5.3
13
17
1.3
0.9
2.5
0.1
0.12
PBS/EtOH (70:30)b
16.90 ± 0.36
9.60 ± 0.21
11.86 ± 0.23
5.10 ± 0.16
10.10 ± 0.22
16.30 ± 0.25
1.28 ± 0.05
0.47 ± 0.01
0.72 ± 0.02
0.42 ± 0.01
0.10 ± 0.003
Clonidine
Imipramine
Testosterone
Caffeine
Enoxacine
Piroxicam
Norfloxacin
Theophylline
a Taken from Ref.40.
b Data are the mean ± SD of three independent experiments.
Figure 6. Lineal correlation between experimental and reported permeability of commercial drugs
using the PAMPA-BBB assay. Pe (exp.) = 0.9163 × Pe (bibl.) − 0.2247 (R2 = 0.9558).

Table 4. Ranges of permeability of PAMPA-BBB assays (Pe × 10-6 cm/s).
High BBB permeation predicted (CNS +)
Pe > 3.44
Uncertain BBB permeation (CNS +/-)
Low BBB permeation predicted (CNS -)
3.44 > Pe > 1.61
Pe < 1.61
Table 5. Permeability results Pe (× 10-6 cm/s) from the PAMPA-BBB assay for selected
homoisoflavonoid Mannich base derivatives with their predicted penetration into the CNS.
Compd.a
Pe (× 10-6 cm/s)b
11.10 ± 0.48
10.37 ± 0.54
6.05 ± 0.18
Prediction
CNS +
CNS +
CNS +
CNS +
10c
10d
10j
10.46 ± 0.44
10k
a
Compounds were dissolved in DMSO at 5 mg/mL and diluted with PBS/EtOH (70:30). The final
concentration of each compound was 100 μg/mL.
b Data are the mean ± SD of three independent experiments.
2.2.10. Cytotoxic effect on the PC-12 cells
To examine the cytotoxicity of the representative compound 10d, PC-12 cells were incubated
with 10d at three different concentrations (1, 10 and 100 μM) and the cell viability was tested using
the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. The active
mitochondrias in living cells can cleave MTT to produce formazan, the amount of which is a direct
reflection of the living cell number. As shown in Figure 7, compound 10d showed no influence on
the cell viability at the concentrations of 1 and 10 μM. With increased concentration to 100 μM, 10d
induced a decrease of cell viability (63.4%). These results revealed that 10d had a wide therapeutic
safety range.

Figure 7. Effect of compounds 10d on the viability of PC-12 cells. Data correspond to the mean ±
SEM of different experiments performed in triplicate.
3. Conclusion
In the current study, we have designed, synthesized and evaluated a series of homoisoflavonoid
Mannich base derivatives as potent multifunctional agents for AD treatment. The in vitro assays
showed that these derivatives exhibited selectively inhibitory activities toward AChE and MAO-B
with good multifunctional properties. Among them, compound 10d and 10k displayed the
comprehensive advantages, with excellent AChE inhibitory activities (IC50 = 0.00249 ± 0.00008 μM
and 18.1 ± 0.882 μM, respectively), MAO-B inhibitory activities (IC50 = 1.74 ± 0.0581 μM and 1.34
± 0.0772 μM, respectively), good self- and Cu2+-induced Aβ1-42 aggregation inhibitory potency,
antioxidant activities, biometal chelating abilities and high BBB permeabilities. Collectively, these
multifunctional properties make 10d and 10k excellent candidates for the development of efficient
drugs against AD. Further studies based on these results are in progress.
4. Experimental section
4.1. Chemistry
1
The H NMR and 13C NMR spectra were recorded using TMS as the internal standard on a
Varian INOVA spectrometer. Coupling constants are given in Hz. MS spectra were recorded on a
Agilent-6210 TOF LC-MS Spectrometer. Melting points were determined on a YRT-3 melting-point
apparatus (China) and are uncorrected. Reactions were monitored by thin-layer chromatography
(TLC) using silica gel GF254 plates from Qingdao Haiyang Chemical Co. Ltd. (China) and the spots

were visualized by UV light (254 nm). Column chromatography was performed using silica gel
(230-400 mesh) purchased from Qingdao Haiyang Chemical Co. Ltd. (China).
4.1.1. General procedure for the synthesis of 3a-k, 4a-e and 5
The mixture of secondary amine (8.19 mmol) and paraformaldehyde (8.19 mmol) in ethanol (7
mL) was refluxed for 1 h, until paraformaldehyde was dissolved. Then compound 1 or 2 (500 mg,
4.09 mmol) was added in one portion. The reaction mixture was refluxed for 24-48 h. Ethanol was
removed under reduced pressure. The residue was diluted with water and acidified with an aqueous
solution of hydrochloric acid (10%). The aqueous phase was washed with ethyl acetate, and then
basified with saturated NaHCO3 solution. The alkaline aqueous phase was again extracted with
CH2Cl2. The combined organic phase washed with brine, dried over Na2SO4, filtered and
concentrated under reduced pressure to afford the corresponding crude Mannich base intermediates
3a-k, 4a-e and 5.
4.1.1.1.4-((dimethylamino)methyl)-3-hydroxybenzaldehyde (3a)
Compound 3a was synthesized from 1 and dimethylamine hydrochloride according to the
general procedure. After purification by chromatography on silica gel using petroleum ether/acetone
(5:1) as eluent, the pure product 3a was obtained as a yellow oil, yield 27.0%.
4.1.1.2. 4-((diethylamino)methyl)-3-hydroxybenzaldehyde (3b) and 2-((diethylamino)methyl)-3
-hydroxybenzaldehyde (4a)
Compound 3b and 4a were synthesized from 1 and diethylamine according to the general
procedure. After purification by chromatography on silica gel using petroleum ether/acetone (20:1)
as eluent, the pure product 3b and 4a were obtained as yellow oil, yield 28.5% and 26.3%
respectively.
4.1.1.3.3-hydroxy-4-(pyrrolidin-1-ylmethyl)benzaldehyde (3c)
Compound 3c was synthesized from 1 and pyrrolidine according to the general procedure. After
purification by chromatography on silica gel using petroleum ether/acetone (15:1) as eluent, the pure
product 3c was obtained as a yellow oil, yield 33.1%.
4.1.1.4.3-hydroxy-4-(piperidin-1-ylmethyl)benzaldehyde (3d)
Compound 3d was synthesized from 1 and piperidine according to the general procedure. After
purification by chromatography on silica gel using petroleum ether/acetone (20:1) as eluent, the pure
product 3d was obtained as a yellow oil, yield 63.4%.

4.1.1.5. 3-hydroxy-4-(morpholinomethyl)benzaldehyde (3e) and 3-hydroxy-2-(morpholino
methyl)benzaldehyde (4b)
Compound 3e and 4b were synthesized from 1 and morpholine according to the general
procedure. After purification by chromatography on silica gel using petroleum ether/ethyl acetate
(15:1) as eluent, the pure product 3e and 4b was obtained as yellow oil, yield 35.5% and 19.7%
respectively.
4.1.1.6. 3-hydroxy-4-((4-methylpiperazin-1-yl)methyl)benzaldehyde (3f) and 3-hydroxy-2-((4-
methylpiperazin-1-yl)methyl)benzaldehyde (4c)
Compound 3f and 4c were synthesized from 1 and 1-methylpiperazine according to the general
procedure. After purification by chromatography on silica gel using petroleum ether/acetone (20:1)
as eluent, the pure product 3f and 4c were obtained as yellow oil, yield 43.6% and 15.4%
respectively.
4.1.1.7. 4-((benzyl(methyl)amino)methyl)-3-hydroxybenzaldehyde (3g)
Compound 3g was synthesized from 1 and N-methyl-1-phenylmethanamine according to the
general procedure. After purification by chromatography on silica gel using petroleum ether/ethyl
acetate (40:1) as eluent, the pure product 3g was obtained as a yellow oil, yield 27.0%.
4.1.1.8. 3-hydroxy-4-(((2-methoxybenzyl)(methyl)amino)methyl)benzaldehyde
3-hydroxy-2-(((2-methoxybenzyl)(methyl)amino)methyl)benzaldehyde (4d)
(3h)
and
Compound 3h and 4d were synthesized from
1
and 1-(2-methoxyphenyl)-N-
methylmethanamine according to the general procedure. After purification by chromatography on
silica gel using dichloromethane/ethyl acetate (40:1) as eluent, the pure product 3h and 4d were
obtained as yellow oil, yield 28.2% and 17.3% respectively.
4.1.1.9. 4-((benzyl(ethyl)amino)methyl)-3-hydroxybenzaldehyde (3i) and 2-((benzyl(ethyl)
amino)methyl)-3-hydroxybenzaldehyde (4e)
Compound 3i and 4e were synthesized from 1 and N-benzylethanamine according to the general
procedure. After purification by chromatography on silica gel using petroleum ether/ethyl acetate
(80:1) as eluent, the pure product 3i and 4e were obtained as yellow oil, yield 31.5% and 13.2%
respectively.
4.1.1.10.4-((ethyl(2-methoxybenzyl)amino)methyl)-3-hydroxybenzaldehyde (3j)
Compound 3j was synthesized from 1 and N-(2-methoxybenzyl)ethanamine according to the

general procedure. After purification by chromatography on silica gel using dichloromethane as
eluent, the pure product 3j was obtained as a yellow oil, yield 22.8%.
4.1.1.11. 4-(((2-(dimethylamino)benzyl)(ethyl)amino)methyl)-3-hydroxybenzaldehyde (3k)
Compound 3k was synthesized from 1 and 2-((ethylamino)methyl)-N,N-dimethylaniline
according to the general procedure. After purification by chromatography on silica gel using
petroleum ether/acetone (40:1) as eluent, the pure product 3k was obtained as a yellow oil, yield
67.6%.
4.1.1.12. 4-hydroxy-3-(piperidin-1-ylmethyl)benzaldehyde (5)
Compound 5 was synthesized from 2 and piperidine according to the general procedure. After
purification by chromatography on silica gel using dichloromethane/acetone (30:1) as eluent, the
pure product 5 was obtained as a yellow oil, yield 35.7%.
4.1.2. 3-chloro-1-(2-hydroxy-4,5-dimethoxyphenyl)propan-1-one (7)
The borontrifluoride etherate (3.93 mL, 31.2 mmol) was added dropwise over 5 min to a
mixture of compound 6 (4.00 g, 26.0 mmol) and 3-chloropropanoyl chloride (4.96 mL, 52.0 mmol).
The mixture was kept at 60 °C for 5 h and then quenched by pouring into the ice-cold water (80 mL).
The product was extracted with CH2Cl2 (4 × 50 mL), then washed with brine, dried over Na2SO4,
filtered and concentrated under reduced pressure to give compound 7 (8.25 g) as a reddish brown oil,
which was used without further purification.
4.1.3. 6,7-dimethoxychroman-4-one (8)
A mixture of crude compound 7 (8.25 g) and K2CO3 (9.32 g) in ethanol (300 mL) was stirred
overnight at room temperature. The reaction mixture was filtered, and the filtrate was concentrated
under reduced pressure. The residue was then diluted with ethyl acetate (80 mL) and water (80 mL).
The aqueous phase was extracted with ethyl acetate (4 × 70 mL). The combined organic phase was
washed with 5% NaHCO3 solution and brine, dried over Na2SO4, filtered and concentrated under
reduced pressure to give compound 8 (5.45 g) as a brown oil. After purification by column
chromatography on silica gel using petroleum ether/acetone (15:1) as eluent, the pure product 8 was
obtained as an off-white solid, yield 38.5% (2.08 g, two steps), mp 120-121 °C (Lit.43 120-121 °C).
4.1.4. General procedure for the synthesis of 9, 10a-k and 11a-e
To a solution of 8 (1.0 equiv.) in ethanol were added an aqueous solution of potassium
hydroxide (50%, 10 equiv.) and the corresponding substituted benzaldehyde (1.5 equiv., 1, 3a-k and

4a-e). The solution was stirred overnight under argon atmosphere and then ethanol was removed
under reduced pressure. The mixture was neutralized with acetic acid and then basified with
saturated NaHCO3 solution to adjust the pH to 7-8. The aqueous phase was extracted with CH2Cl2,
then washed with brine, dried over Na2SO4, filtered and concentrated under reduced pressure to give
the crude compound 9, 10a-k and 11a-e.
4.1.4.1.(E)-3-(3-hydroxybenzylidene)-6,7-dimethoxychroman-4-one (9)
Compound 9 was synthesized from 8 and 1 according to the general procedure. After
purification by chromatography on silica gel using petroleum ether/ethyl acetate (2:1) as eluent, the
1
pure product 9 was obtained as a yellow solid, yield 18.0%, mp 191-194 °C. H NMR (400 MHz,
DMSO-d6) δ 9.70 (s, 1H), 7.60 (s, 1H), 7.30-7.22 (m, 2H), 7.15 (s, 1H), 6.85-6.80 (m, 2H), 6.64 (s,
1H), 5.35 (s, 2H), 3.84 (s, 3H), 3.78 (s, 3H).
4.1.4.2. (E)-3-(4-((dimethylamino)methyl)-3-hydroxybenzylidene)-6,7-dimethoxychroman-4-on
e (10a)
Compound 10a was synthesized from 8 and 3a according to the general procedure. After
purification by chromatography on silica gel using petroleum ether/ethyl acetate (1:2) as eluent, the
pure product 10a was obtained as a yellow solid, yield 20.3%, mp 132-134 °C. 1H NMR (400 MHz,
CDCl3) δ 7.76 (s, 1H), 7.40 (s, 1H), 7.02 (d, J = 8.0 Hz, 1H), 6.78-6.70 (m, 2H), 6.42 (s, 1H), 5.34 (s,
2H), 3.91 (s, 6H), 3.69 (s, 2H), 2.36 (s, 6H). 13C NMR (100 MHz, CDCl3) δ 180.8, 158.1, 157.6,
156.1, 144.7, 136.5, 135.0, 130.5, 128.5, 123.1, 121.0, 117.0, 114.1, 107.3, 100.0, 68.0, 62.4, 56.2,
56.1, 44.4 (2C). HR-ESI-MS: Calcd. for C21H24NO5 [M + H]+ : 370.1654, found : 370.1661.
4.1.4.3. (E)-3-(4-((diethylamino)methyl)-3-hydroxybenzylidene)-6,7-dimethoxychroman-4-one
(10b)
Compound 10b was synthesized from 8 and 3b according to the general procedure. After
purification by chromatography on silica gel using petroleum ether/ethyl acetate (2:1) as eluent, the
pure product 10b was obtained as a yellow gum, yield 20.0%. 1H NMR (400 MHz, CDCl3) δ 7.77 (s,
1H), 7.41 (s, 1H), 7.03 (d, J = 7.6 Hz, 1H), 6.74 (d, J = 7.6 Hz, 1H), 6.73 (s, 1H), 6.43 (s, 1H), 5.35
(d, J = 1.6 Hz, 2H), 3.92 (s, 6H), 3.82 (s, 2H), 2.66 (q, J = 7.2 Hz, 4H), 1.14 (t, J = 7.2 Hz, 6H).
HR-ESI-MS: Calcd. for C23H28NO5 [M + H]+ : 398.1967, found : 398.1968.
4.1.4.4. (E)-3-(3-hydroxy-4-(pyrrolidin-1-ylmethyl)benzylidene)-6,7-dimethoxychroman-4-one
(10c)

Compound 10c was synthesized from 8 and 3c according to the general procedure. After
purification by chromatography on silica gel using petroleum ether/ethyl acetate (1:1) as eluent, the
1
pure product 10c was obtained as a yellow solid, yield 20.4%, mp 80-82 °C. H NMR (400 MHz,
CDCl3) δ 7.77 (s, 1H), 7.41 (s, 1H), 7.04 (d, J = 8.0 Hz, 1H), 6.74 (d, J = 8.0 Hz, 1H), 6.74 (s, 1H),
6.43 (s, 1H), 5.35 (d, J = 1.6 Hz, 2H), 3.92 (s, 6H), 3.88 (s, 2H), 2.68 (brs, 4H), 1.92-1.84 (m, 4H).
13C NMR (100 MHz, DMSO-d6) δ 179.8, 157.4, 157.1, 156.4, 144.9, 135.8, 134.1, 130.6, 129.2,
125.5, 121.2, 116.7, 113.7, 107.2, 100.8, 67.9, 56.4 (2C), 55.9, 53.4 (2C), 23.5 (2C). HR-ESI-MS:
Calcd. for C23H26NO5 [M + H]+ : 396.1811, found : 396.1816.
4.1.4.5. (E)-3-(3-hydroxy-4-(piperidin-1-ylmethyl)benzylidene)-6,7-dimethoxychroman-4-one
(10d)
Compound 10d was synthesized from 8 and 3d according to the general procedure. After
purification by chromatography on silica gel using petroleum ether/ethyl acetate (2:1) as eluent, the
pure product 10d was obtained as a yellow solid, yield 20.2%, mp 146-147 °C. 1H NMR (400 MHz,
CDCl3) δ 7.77 (s, 1H), 7.41 (s, 1H), 7.03 (d, J = 8.0 Hz, 1H), 6.75 (s, 1H), 6.74 (d, J = 8.0 Hz, 1H),
6.44 (s, 1H), 5.34 (d, J = 1.6 Hz, 2H), 3.92 (s, 6H), 3.74 (s, 2H), 2.63 (brs, 4H), 1.69-1.66 (m, 4H),
1.52 (brs, 2H). 13C NMR (100 MHz, CDCl3) δ 180.9, 158.1, 157.7, 156.1, 144.8, 136.6, 135.1, 130.6,
128.9, 122.8, 121.1, 117.2, 114.2, 107.4, 100.1, 68.1, 61.6, 56.3, 56.2, 53.9 (2C), 25.6 (2C), 23.8.
HR-ESI-MS: Calcd. for C24H28NO5 [M + H]+ : 410.1967, found : 410.1965.
4.1.4.6. (E)-3-(3-hydroxy-4-(morpholinomethyl)benzylidene)-6,7-dimethoxychroman-4-one
(10e)
Compound 10e was synthesized from 8 and 3e according to the general procedure. After
purification by chromatography on silica gel using petroleum ether/ethyl acetate (2:1) as eluent, the
1
pure product 10e was obtained as a yellow solid, yield 25.2%, mp 86-88 °C. H NMR (400 MHz,
CDCl3) δ 7.77 (s, 1H), 7.41 (s, 1H), 7.07 (d, J = 8.4 Hz, 1H), 6.76 (d, J = 8.4 Hz, 1H), 6.76 (s, 1H),
13
6.43 (s, 1H), 5.33 (d, J = 1.6 Hz, 2H), 3.92 (s, 6H), 3.92 (s, 6H), 3.77 (s, 6H), 2.62 (brs, 4H). C
NMR (100 MHz, CDCl3) δ 180.8, 157.6, 157.5, 156.1, 144.8, 136.3, 135.3, 130.7, 129.0, 121.9,
121.3, 117.1, 114.1, 107.3, 100.0, 68.0, 66.6 (2C), 61.5, 56.2, 56.1, 52.8 (2C). HR-ESI-MS: Calcd.
for C23H26NO6 [M + H]+ : 412.1760, found : 412.1759.
4.1.4.7. (E)-3-(3-hydroxy-4-((4-methylpiperazin-1-yl)methyl)benzylidene)-6,7-dimethoxychrom
an-4-one (10f)

Compound 10f was synthesized from 8 and 3f according to the general procedure. After
purification by chromatography on silica gel using petroleum ether/acetone (3:1) as eluent, the pure
product 10f was obtained as a yellow solid, yield 20.8%, mp 85-88 °C. 1H NMR (400 MHz, CDCl3)
δ 7.77 (s, 1H), 7.41 (s, 1H), 7.05 (d, J = 8.0 Hz, 1H), 6.75 (d, J = 8.0 Hz, 1H), 6.74 (s, 1H), 6.43 (s,
13
1H), 5.33 (d, J = 1.2 Hz, 2H), 3.92 (s, 6H), 3.77 (s, 2H), 2.62 (brs, 8H), 2.35 (s, 3H). C NMR (100
MHz, CDCl3) δ 180.8, 157.7, 157.6, 156.1, 144.8, 136.4, 135.2, 130.6, 128.8, 122.4, 121.2, 117.1,
114.1, 107.3, 100.0, 68.0, 61.0, 56.2, 56.1, 54.7 (2C), 52.3 (2C), 45.7. HR-ESI-MS: Calcd. for
C24H29N2O5 [M + H]+ : 425.2076, found : 425.2074.
4.1.4.8. (E)-3-(4-((benzyl(methyl)amino)methyl)-3-hydroxybenzylidene)-6,7-dimethoxychroma
n-4-one (10g)
Compound 10g was synthesized from 8 and 3g according to the general procedure. After
purification by chromatography on silica gel using petroleum ether/ethyl acetate (4:1) as eluent, the
pure product 10g was obtained as a yellow solid, yield 22.9%, mp 185-187 °C. 1H NMR (400 MHz,
CDCl3) δ 7.78 (s, 1H), 7.41 (s, 1H), 7.39-7.31 (m, 5H), 7.06 (d, J = 7.6 Hz, 1H), 6.78 (s, 1H), 6.76 (d,
J = 7.6 Hz, 1H), 6.43 (s, 1H), 5.34 (d, J = 1.6 Hz, 2H), 3.92 (s, 6H), 3.81 (s, 2H), 3.65 (s, 2H), 2.29 (s,
13
3H). C NMR (100 MHz, CDCl3) δ 180.9, 157.8, 157.7, 156.1, 144.8, 136.5, 136.3, 135.2, 130.6,
129.4 (2C), 128.8, 128.7 (2C), 127.9, 123.1, 121.3, 117.2, 114.2, 107.4, 100.0, 68.1, 61.4, 60.4, 56.3,
56.2, 41.3. HR-ESI-MS: Calcd. for C27H28NO5 [M + H]+ : 446.1967, found : 446.1962.
4.1.4.9. (E)-3-(3-hydroxy-4-(((2-methoxybenzyl)(methyl)amino)methyl)benzylidene)-6,7-dimet
hoxychroman-4-one (10h)
Compound 10h was synthesized from 8 and 3h according to the general procedure. After
purification by chromatography on silica gel using dichloromethane/methanol (60:1) as eluent, the
1
pure product 10h was obtained as a yellow solid, yield 33.0%, mp 80-82 °C. H NMR (400 MHz,
CDCl3) δ 7.76 (s, 1H), 7.41 (s, 1H), 7.31 (t, J = 8.0 Hz, 1H), 7.22 (d, J = 6.4 Hz, 1H), 7.05 (d, J = 8.0
Hz, 1H), 6.95-6.91 (m, 2H), 6.75 (s, 1H), 6.74 (d, J = 8.0 Hz, 1H), 6.42 (s, 1H), 5.33 (d, J = 1.6 Hz,
13
2H), 3.92 (s, 6H), 3.89 (s, 3H), 3.79 (s, 2H), 3.70 (s, 2H), 2.23 (s, 3H). C NMR (100 MHz, CDCl3)
δ 180.8, 158.1, 158.0, 157.6, 156.0, 144.7, 136.6, 134.9, 131.3, 130.4, 129.4, 128.7, 124.5, 123.7,
121.0, 120.2, 117.0, 114.1, 110.5, 107.3, 100.0, 68.1, 60.2, 57.0, 56.2, 56.1, 55.1, 40.9. HR-ESI-MS:
Calcd. for C28H30NO6 [M + H]+ : 476.2073, found : 476.2081.
4.1.4.10.(E)-3-(4-((benzyl(ethyl)amino)methyl)-3-hydroxybenzylidene)-6,7-dimethoxychroman-

4-one (10i)
Compound 10i was synthesized from 8 and 3i according to the general procedure. After
purification by chromatography on silica gel using dichloromethane/acetone (80:1) as eluent, the
1
pure product 10i was obtained as a yellow solid, yield 21.9%, mp 68-70 °C. H NMR (400 MHz,
CDCl3) δ 11.4 (brs, 1H), 7.77 (s, 1H), 7.41 (s, 1H), 7.39-7.30 (m, 5H), 7.06 (d, J = 8.0 Hz, 1H), 6.76
(s, 1H), 6.75 (d, J = 8.0 Hz, 1H), 6.42 (s, 1H), 5.34 (d, J = 1.6 Hz, 2H), 3.92 (s, 6H), 3.83 (s, 2H),
3.69 (s, 2H), 2.64 (q, J = 7.2 Hz, 2H), 1.18 (t, J = 7.2 Hz, 3H). 13C NMR (100 MHz, CDCl3) δ 180.8,
157.9, 157.6, 156.1, 144.8, 136.5, 136.4, 135.0, 130.6, 129.5 (2C), 128.8, 128.6 (2C), 127.7, 123.4,
121.2, 117.1, 114.2, 107.4, 100.0, 68.1, 57.5, 56.6, 56.3, 56.2, 46.7, 11.0. HR-ESI-MS: Calcd. for
C28H30NO5 [M + H]+ : 460.2124, found : 460.2123.
4.1.4.11.(E)-3-(4-((ethyl(2-methoxybenzyl)amino)methyl)-3-hydroxybenzylidene)-6,7-dimethox
ychroman-4-one (10j)
Compound 10j was synthesized from 8 and 3j according to the general procedure. After
purification by chromatography on silica gel using dichloromethane/acetone (50:1) as eluent, the
1
pure product 10j was obtained as a yellow solid, yield 20.2%, mp 81-83 °C. H NMR (400 MHz,
CDCl3) δ 7.75 (s, 1H), 7.40 (s, 1H), 7.28 (t, J = 8.0 Hz, 1H), 7.23 (d, J = 6.4 Hz, 1H), 7.02 (d, J = 8.0
Hz, 1H), 6.95-6.90 (m, 2H), 6.72 (d, J = 8.0 Hz, 1H), 6.10 (s, 1H), 6.42 (s, 1H), 5.32 (d, J = 1.6 Hz,
2H), 3.91 (s, 6H), 3.88 (s, 3H), 3.80 (s, 2H), 3.73 (s, 2H), 2.62 (q, J = 7.2 Hz, 2H), 1.16 (t, J = 7.2 Hz,
3H). 13C NMR (100 MHz, CDCl3) δ 180.9, 158.1 (2C), 157.6, 156.1, 144.8, 136.7, 134.8, 131.4,
130.4, 129.3, 128.7, 124.7, 123.9, 121.0, 120.3, 117.0, 114.2, 110.6, 107.4, 100.0, 68.1, 56.4, 56.2
(2C), 55.2, 52.8, 46.8, 10.6. HR-ESI-MS: Calcd. for C29H32NO6 [M + H]+ : 490.2230, found :
490.2233.
4.1.4.12.(E)-3-(4-(((2-(dimethylamino)benzyl)(ethyl)amino)methyl)-3-hydroxybenzylidene)-6,7-
dimethoxychroman-4-one (10k)
Compound 10k was synthesized from 8 and 3k according to the general procedure. After
purification by chromatography on silica gel using petroleum ether/acetone (5:1) as eluent, the pure
product 10k was obtained as a yellow solid, yield 15.4%, mp 80-82 °C. 1H NMR (400 MHz, CDCl3)
δ 7.76 (s, 1H), 7.40 (s, 1H), 7.36 (d, J = 7.2 Hz, 1H), 7.26 (t, J = 7.2 Hz, 1H), 7.16 (d, J = 7.2 Hz,
1H), 7.09 (t, J = 7.2 Hz, 1H), 7.04 (d, J = 7.6 Hz, 1H), 6.75 (s, 1H), 6.73 (d, J = 7.6 Hz, 1H), 6.42 (s,
1H), 5.33 (d, J = 1.6 Hz, 2H), 3.91 (s, 6H), 3.81 (s, 4H), 2.67 (s, 6H), 2.59 (q, J = 6.8 Hz, 2H), 1.15 (t,

J = 6.8 Hz, 3H). 13C NMR (100 MHz, CDCl3) δ 180.9, 158.0, 157.6, 156.1, 153.5, 144.8, 136.6,
134.9, 131.4, 130.6, 130.4, 128.8, 128.4, 123.8, 123.7, 121.0, 119.8, 117.0, 114.2, 107.4, 100.0, 68.1,
56.9, 56.2 (2C), 52.9, 46.8, 45.2 (2C), 10.7. HR-ESI-MS: Calcd. for C30H35N2O5 [M + H]+ :
503.2546, found : 503.2549.
4.1.4.13. (E)-3-(2-((diethylamino)methyl)-3-hydroxybenzylidene)-6,7-dimethoxychroman-4-one
(11a)
Compound 11a was synthesized from 8 and 4a according to the general procedure. After
purification by chromatography on silica gel using petroleum ether/ethyl acetate (2:1) as eluent, the
1
pure product 11a was obtained as a yellow solid, yield 21.2%, mp 72-75 °C. H NMR (400 MHz,
CDCl3) δ 7.85 (s, 1H), 7.42 (s, 1H), 7.18 (t, J = 8.0 Hz, 1H), 6.86 (d, J = 8.0 Hz, 1H), 6.74 (d, J = 8.0
Hz, 1H), 6.42 (s, 1H), 5.11 (s, 2H), 3.92 (s, 6H), 3.82 (s, 2H), 2.64 (q, J = 7.2 Hz, 4H), 1.12 (t, J =
7.2 Hz, 6H). HR-ESI-MS: Calcd. for C23H28NO5 [M + H]+ : 398.1967, found : 398.1962.
4.1.4.14.(E)-3-(3-hydroxy-2-(morpholinomethyl)benzylidene)-6,7-dimethoxychroman-4-one
(11b)
Compound 11b was synthesized from 8 and 4b according to the general procedure. After
purification by chromatography on silica gel using petroleum ether/ethyl acetate (2:1) as eluent, the
pure product 11b was obtained as a yellow solid, yield 16.7%, mp 193-195 °C. 1H NMR (400 MHz,
CDCl3) δ 11.4 (brs, 1H), 7.84 (s, 1H), 7.42 (s, 1H), 7.21 (t, J = 8.0 Hz, 1H), 6.89 (d, J = 8.0 Hz, 1H),
6.51 (d, J = 8.0 Hz, 1H), 6.42 (s, 1H), 5.08 (s, 2H), 3.92 (s, 6H), 3.77 (s, 6H), 2.60 (brs, 4H). 13C
NMR (100 MHz, CDCl3) δ 180.9, 158.5, 158.0, 156.4, 145.0, 135.1, 134.1, 132.4, 128.6, 120.2,
119.4, 117.6, 114.1, 107.3, 100.1, 67.9, 66.6 (2C), 57.7, 56.3, 56.2, 52.7 (2C). HR-ESI-MS: Calcd.
for C23H26NO6 [M + H]+ : 412.1760, found : 412.1768.
4.1.4.15.(E)-3-(3-hydroxy-2-((4-methylpiperazin-1-yl)methyl)benzylidene)-6,7-dimethoxychrom
an-4-one (11c)
Compound 11c was synthesized from 8 and 4c according to the general procedure. After
purification by chromatography on silica gel using petroleum ether/acetone (3:1) as eluent, the pure
1
product 11c was obtained as a yellow solid, yield 27.0%, mp 165-167 °C. H NMR (400 MHz,
CDCl3) δ 7.84 (s, 1H), 7.42 (s, 1H), 7.19 (t, J = 8.0 Hz, 1H), 6.87 (d, J = 8.0 Hz, 1H), 6.49 (d, J = 8.0
Hz, 1H), 6.42 (s, 1H), 5.08 (s, 2H), 3.92 (s, 6H), 3.77 (s, 2H), 2.59 (brs, 8H), 2.33 (s, 3H).
HR-ESI-MS: Calcd. for C24H29N2O5 [M + H]+ : 425.2076, found : 425.2070.

4.1.4.16.(E)-3-(3-hydroxy-2-(((2-methoxybenzyl)(methyl)amino)methyl)benzylidene)-6,7-dimet
hoxychroman-4-one (11d)
Compound 11d was synthesized from 8 and 4d according to the general procedure. After
purification by chromatography on silica gel using dichloromethane/methanol (60:1) as eluent, the
1
pure product 11d was obtained as a yellow solid, yield 13.1%, mp 78-80 °C. H NMR (400 MHz,
CDCl3) δ 7.89 (s, 1H), 7.43 (s, 1H), 7.29 (t, J = 8.0 Hz, 1H), 7.20-7.15 (m, 2H), 6.93-6.86 (m, 3H),
6.47 (d, J = 7.2 Hz, 1H), 6.42 (s, 1H), 5.09 (d, J = 1.2 Hz, 2H), 3.94 (s, 3H), 3.92 (s, 3H), 3.88 (s,
3H), 3.78 (s, 2H), 3.66 (s, 2H), 2.22 (s, 3H). 13C NMR (100 MHz, CDCl3) δ 181.0, 158.9, 158.1,
158.0, 156.3, 144.9, 135.6, 133.7, 132.0, 131.4, 129.4, 128.2, 124.6, 121.1, 120.2, 119.7, 117.5,
114.1, 110.5, 107.3, 100.1, 68.0, 57.0, 56.4, 56.3, 56.2, 55.2, 40.8. HR-ESI-MS: Calcd. for
C28H30NO6 [M + H]+ : 476.2073, found : 476.2075.
4.1.4.17.(E)-3-(2-((benzyl(ethyl)amino)methyl)-3-hydroxybenzylidene)-6,7-dimethoxychroman-
4-one (11e)
Compound 11e was synthesized from 8 and 4e according to the general procedure. After
purification by chromatography on silica gel using dichloromethane/acetone (80:1) as eluent, the
1
pure product 11e was obtained as a yellow solid, yield 13.6%, mp 63-65 °C. H NMR (400 MHz,
CDCl3) δ 12.0 (brs, 1H), 7.86 (s, 1H), 7.42 (s, 1H), 7.36-7.26 (m, 5H), 7.19 (t, J = 7.2 Hz, 1H), 6.88
(d, J = 7.2 Hz, 1H), 6.48 (d, J = 7.2 Hz, 1H), 6.42 (s, 1H), 5.10 (d, J = 1.2 Hz, 2H), 3.93 (s, 3H), 3.92
(s, 3H), 3.82 (s, 2H), 3.65 (s, 2H), 2.62 (q, J = 7.2 Hz, 2H), 1.16 (t, J = 7.2 Hz, 3H). 13C NMR (100
MHz, CDCl3) δ 181.0, 158.8, 158.0, 156.4, 145.0, 136.5, 135.4, 133.8, 132.1, 129.4 (2C), 128.6 (2C),
128.2, 127.7, 120.8, 120.0, 117.6, 114.1, 107.3, 100.1, 68.0, 57.7, 56.3, 56.2, 52.9, 46.9, 11.2.
HR-ESI-MS: Calcd. for C28H30NO5 [M + H]+ : 460.2124, found : 460.2127.
4.1.5. (E)-3-(4-hydroxy-3-(piperidin-1-ylmethyl)benzylidene)-6,7-dimethoxychroman-4-one
(12)
A mixture of compound 8 (50.0 mg, 0.240 mmol) and 5 (80.0 mg, 36.5 mmol) in 85%
phosphoric acid (1 mL) was heated at 80 °C for 10 h under argon atmosphere. After cooling, the
mixture was basified with saturated NaHCO3 solution to adjust the pH to 7-8. The aqueous phase
was extracted with CH2Cl2, then washed with brine, dried over Na2SO4, filtered and concentrated
under reduced pressure to give the crude product as a reddish brown oil. After purification by
chromatography on silica gel using petroleum ether/acetone (3:1) as eluent, the pure product 12 (50.0

1
mg) was obtained as a yellow solid, yield 51.0%, mp 78-80 °C. H NMR (400 MHz, CDCl3) δ 7.76
(s, 1H), 7.41 (s, 1H), 7.13 (d, J = 8.4 Hz, 1H), 6.98 (s, 1H), 6.87 (d, J = 8.4 Hz, 1H), 6.42 (s, 1H),
13
5.35 (s, 2H), 3.92 (s, 6H), 3.72 (s, 2H), 2.54 (brs, 4H), 1.69-1.60 (m, 4H), 1.51 (brs, 2H). C NMR
(100 MHz, CDCl3) δ 180.8, 159.8, 157.3, 155.8, 144.7, 136.8, 131.1, 130.9, 127.9, 125.5, 121.8,
116.4, 114.2, 107.4, 99.9, 68.1, 61.8, 56.2, 56.1, 53.8 (2C), 25.7 (2C), 23.8. HR-ESI-MS: Calcd. for
C24H28NO5 [M + H]+ : 410.1967, found : 410.1965.
4.2. Pharmacology
4.2.1. Inhibition experiments of AChE and BuChE
To evaluate the cholinesterase activities of the compounds, the modified Ellman’s method44 was
performed, using purified AChE from Electrophorus electricus (Sigma-Aldrich Co.) and BuChE
from rat serum. For EeAChE inhibition assay, the reaction mixture (100 µL) consisted of
acetylthiocholine iodide (1 mmol/L, 30 µL) (J&K Scientific), phosphate-buffered solution (0.1
mmol/L, pH = 8.00, 40 µL), EeAChE (0.05 U/mL, final concentration, 10 µL) and different
concentrations of test compounds (20 µL). After the addition of 5,5’-dithiobis-2-nitrobenzoic acid
(DTNB, 0.2%, 30 µL) (J&K Scientific), the mixture was incubated at 37 °C for 15 min and then
detected at 412 nm in a Varioskan Flash Multimode Reader (Thermo Scientific). As for the BuChE
inhibition assay, 25% rat serum (10 µL) was used and the assay was carried out in a phosphate buffer
(0.1 mmol/L, pH = 7.40). Changes in absorbance were detected at 405 nm. The IC50 values were
calculated as the concentration of compound that produces 50% AChE or BuChE activity inhibition.
Donepezil was used as the positive drug. All samples were carried out in triplicate.
4.2.2. Kinetic study for the inhibition of AChE
Kinetic study for the inhibition of AChE was conducted according to a reported method using
purified EeAChE.45 The mixture containing 0.1 M phosphate buffer (pH = 8.00), EeAChE (0.5
U/mL), 0.2% DTNB and different concentrations of test compounds was incubated for 15 min at
37°C, followed by the addition of substrate in different concentrations. Kinetic characterization of
the hydrolysis of ATCh catalyzed by EeAChE was done spectra-metrically at 412 nm. The parallel
control experiment was conducted without inhibitor. The plots were assessed by a weighted least
square analysis that assumed the variance of v to be a constant percentage of v for the entire data set.
Slopes of reciprocal plots were then plotted against the concentration of 10d in a weighted analysis,
and Ki was determined as the intercept on the negative x-axis.

4.2.3. Molecular modeling study of AChE
The simulation systems were built based on the X-ray crystal structures of AChE (PDB code:
1EVE). The original inhibitors and water molecules were removed. The 3D Structure of 10d was
built and performed geometry optimization by molecular mechanics. After addition of Gasteiger
charges, removal of hydrogen atoms, addition of their atomic charges to skeleton atoms, and the
assignment of proper atomic types, the further preparation of the inhibitor was accomplished.
Autotors was then used to define the rotatable bonds in the ligands. Molecular modeling studies were
performed using the AUTODOCK 4.2 program. By using Autodock Tools (ADT; version 1.5.6),
polar hydrogen atoms were added to amino acid residues, and Gasteiger charges were assigned to all
atoms of the enzyme. The resulting enzyme structure was used as an input for the AUTOGRID
program. AUTOGRID performed a pre-calculated atomic affinity grid maps for each atom type in
the ligand, plus an electrostatics map and a separate desolvation map presented in the substrate
molecule. All maps were calculated with 0.375 Å spacing between grid points. The center of the grid
box was placed at the center of donepezil with coordinates x = 2.023, y = 63.295, z = 67.062. The
dimensions of the active site box were set at 60 × 60 × 60 Å. Flexible ligand docking was performed
for the compounds. Each docked system was performed by 100 runs of the AUTODOCK search by
the Lamarckian genetic algorithm (LGA). Other than the referred parameters above, the other
parameters were accepted as default. A cluster analysis was performed on the docking results using a
root mean square (RMS) tolerance of 1.0 and the lowest energy conformation of the highest
populated cluster was selected for analysis. Graphic manipulations and visualizations were done by
Autodock Tools or Discovery Studio 2.5 software.
4.2.4. In vitro antioxidant activity assay
Fluorescein (FL) and 6-hydroxy-2,5,7,8-tetramethylchromane-2- carboxylic acid (Trolox) were
purchased from TCI (Shanghai) Development. 2,2’-Azobis(amidinopropane) dihydrochloride
(AAPH) was purchased from Accela ChemBio Co., Ltd. All the assays were performed with 75 mM
phosphate buffer (pH = 7.40), and the final reaction mixture was 200 μL. Antioxidant (20 μL) and FL
(120 μL, 150 nM final concentration) were added into the wells of a black 96-well plate with Trolox
as a standard (1-8 μM, final concentration). The mixture was incubated for 15 min at 37 °C and then
placed in a Varioskan Flash Multimode Reader (Thermo Scientific). AAPH solution (60 μL, 12 mM
final concentration) was added rapidly and the fluorescence recorded every minute for 90 min with