Design, synthesis, and biological evaluation of novel combretastatin A-4 thio derivatives as microtubule targeting agents

Article abstract of DOI:10.1016/j.ejmech.2017.11.050

A series of novel combretastatin A-4 (CA-4) thio derivatives containing different molecular cores, namely α-phenylcinnamic acids (core 1), (Z)-stilbenes (core 2), 4,5-disubstituted oxazoles (core 3), and 4,5-disubstituted N-methylimidazoles (core 4), as cis-restricted analogues were designed and synthesized. They were selected with the use of a parallel virtual screening protocol including the generation of a virtual combinatorial library based on an elaborated synthesis protocol of CA-4 analogues. The selected compounds were evaluated for antiproliferative activity against a panel of six human cancer cell lines (A431, HeLa, MCF7, MDA-MB-231, A549 and SKOV) and two human non-cancer cell lines (HaCaT and CCD39Lu). Moreover, the effect of the test compounds on the inhibition of tubulin polymerization in vitro was estimated. In the series studied here, oxazole-bridged analogues exhibited the most potent antiproliferative activity. Compounds 23a, 23e, and 23i efficiently inhibited tubulin polymerization with IC₅₀ values of 0.86, 1.05, and 0.85 μM, respectively. Thio derivative 23i, when compared to its oxygen analogue 23j, showed a 5-fold higher inhibitory impact on tubulin polymerization. Compounds 23e and 23i, which showed both best cytotoxic and antitubulin activity, were further studied in terms of their effect on cell cycle distribution and proapoptotic activity. Compound 23e induced a statistically significant block of the cell cycle at the G2/M phase in A431, HaCaT, HeLa, MCF-7, MDA-MB-231, and SKOV-3 cells to an extent comparable to that observed in CA-4. In HeLa and SKOV-3 cells incubated with 23i, a concentration-dependent block of the G2/M phase was observed. The proapoptotic effect of 23e and 23i in A431, HaCaT, MCF-7, MDA-MB-231, and SKOV-3 was demonstrated with ELISA assay and double staining with Annexin V-FITC/PI. The results indicated that compound 23e and 23i may serve as novel lead compounds in research on more effective anticancer agents.

Full text of DOI:10.1016/j.ejmech.2017.11.050

Accepted Manuscript
Design, synthesis, and biological evaluation of novel combretastatin A-4 thio
derivatives as microtubule targeting agents
Tomasz Stefański, Renata Mikstacka, Rafał Kurczab, Zbigniew Dutkiewicz,
Małgorzata Kucińska, Marek Murias, Małgorzata Zielińska-Przyjemska, Michał
Cichocki, Anna Teubert, Mariusz Kaczmarek, Adam Hogendorf, Stanisław Sobiak
PII:
S0223-5234(17)30953-4
10.1016/j.ejmech.2017.11.050
EJMECH 9925
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 27 September 2017
Revised Date: 17 November 2017
Accepted Date: 18 November 2017
Please cite this article as: T. Stefański, R. Mikstacka, Rafał. Kurczab, Z. Dutkiewicz, Mał. Kucińska,
M. Murias, Mał. Zielińska-Przyjemska, Michał. Cichocki, A. Teubert, M. Kaczmarek, A. Hogendorf,
Stanisł. Sobiak, Design, synthesis, and biological evaluation of novel combretastatin A-4 thio derivatives
as microtubule targeting agents, European Journal of Medicinal Chemistry (2018), doi: 10.1016/
j.ejmech.2017.11.050.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT
Design, synthesis, and biological evaluation of novel
combretastatin A-4 thio derivatives as microtubule
targeting agents
1
*
2
3
1
Tomasz Stefański, Renata Mikstacka, Rafał Kurczab, Zbigniew Dutkiewicz, Małgorzata
4
4
5
5
Kucińska, Marek Murias, Małgorzata Zielińska-Przyjemska, Michał Cichocki, Anna
6
7
3
2
Teubert, Mariusz Kaczmarek, Adam Hogendorf, and Stanisław Sobiak
1
Department of Chemical Technology of Drugs, Poznan University of Medical Sciences,
Grunwaldzka 6, 60-780 Poznań, Poland
2
Department of Inorganic and Analytical Chemistry, Collegium Medicum in Bydgoszcz,
Nicolaus Copernicus University in Toruń, Jurasza 2, 85-089 Bydgoszcz, Poland
3
Department of Medicinal Chemistry, Institute of Pharmacology, Polish Academy of Sciences,
Smętna 12, 31-343 Kraków, Poland
4
Department of Toxicology, Poznan University of Medical Sciences, Dojazd 30, 60-631 Poznań,
Poland
5
Department of Pharmaceutical Biochemistry, Poznan University of Medical Sciences,
Święcickiego 4, 60-781 Poznań, Poland
6
Institute of Bioorganic Chemistry, Polish Academy of Sciences, Noskowskiego 12/14, 61-704
Poznań, Poland
1

ACCEPTED MANUSCRIPT
7
Department of Immunology, Poznan University of Medical Sciences, Rokietnicka 5D, 60-806
Poznań, Poland
ABSTRACT: A series of novel combretastatin A-4 (CA-4) thio derivatives containing different
molecular cores, namely α-phenylcinnamic acids (core 1), (Z)-stilbenes (core 2), 4,5-
disubstituted oxazoles (core 3), and 4,5-disubstituted N-methylimidazoles (core 4), as cis-
restricted analogues were designed and synthesized. They were selected with the use of a parallel
virtual screening protocol including the generation of a virtual combinatorial library based on an
elaborated synthesis protocol of CA-4 analogues. The selected compounds were evaluated for
antiproliferative activity against a panel of six human cancer cell lines (A431, HeLa, MCF7,
MDA-MB-231, A549 and SKOV) and two human non-cancer cell lines (HaCaT and CCD39Lu).
Moreover, the effect of the test compounds on the inhibition of tubulin polymerization in vitro
was estimated. In the series studied here, oxazole-bridged analogues exhibited the most potent
antiproliferative activity. Compounds 23a, 23e, and 23i efficiently inhibited tubulin
polymerization with IC values of 0.86, 1.05, and 0.85 µM, respectively. Thio derivative 23i,
5
0
when compared to its oxygen analogue 23j, showed a 5-fold higher inhibitory impact on tubulin
polymerization. Compounds 23e and 23i, which showed both best cytotoxic and antitubulin
activity, were further studied in terms of their effect on cell cycle distribution and proapoptotic
activity. Compound 23e induced a statistically significant block of the cell cycle at the G2/M
phase in A431, HaCaT, HeLa, MCF-7, MDA-MB-231, and SKOV-3 cells to an extent
comparable to that observed in CA-4. In HeLa and SKOV-3 cells incubated with 23i, a
concentration-dependent block of the G2/M phase was observed. The proapoptotic effect of 23e
and 23i in A431, HaCaT, MCF-7, MDA-MB-231, and SKOV-3 was demonstrated with ELISA
2

ACCEPTED MANUSCRIPT
assay and double staining with Annexin V-FITC/PI. The results indicated that compound 23e
and 23i may serve as novel lead compounds in research on more effective anticancer agents.
Keywords: anticancer agents; tubulin polymerization inhibitors; combretastatin A-4; virtual
screening; synthesis.
1
. Introduction
Microtubules are highly dynamic components of the cytoskeleton that consist of αβ-tubulin
heterodimers and are involved in a wide range of various cellular functions, including
maintenance of the cell structure, motility, intercellular transport, and cell division, where they
are responsible for mitotic spindle formation and proper chromosomal separation [1–4]. The
biological importance of microtubules in mitosis and cell division as well as their dynamic nature
makes them a significant and intensively investigated molecular target for anticancer drugs [5–
1
1].
Microtubule-interfering agents (MIAs), of which several are natural products, act via two
different mechanisms of action leading to either the inhibition or enhancement of tubulin
polymerization. Both effects impair microtubule dynamics and have an impact on cell
proliferation. MIAs bind to one of the three major different binding sites on tubulin, i.e. the taxol
binding site, colchicine-binding site, and vinca alkaloid domain [5,6]. Ligands which stimulate
microtubule polymerization, known as microtubule stabilizers such as paclitaxel or epothilone,
bind to the taxol binding site [12,13]. In turn, microtubule destabilizers (tubulin polymerization
inhibitors) induce depolymerization of microtubules. In this group, different vinca alkaloids can
be distinguished, such as vinblastine and vincristine which bind to the vinca domain [14,15], and
3

ACCEPTED MANUSCRIPT
a large class of ligands which interact with the colchicine-binding site, including colchicine (2),
podophyllotoxin (3), steganacin (4), and combretastatins [16–20].
Combretastatins are a group of natural cis-stilbenes that were isolated from the bark of the
African bush willow tree Combretum caffrum over 30 years ago [21–24]. The most active
member of this group as described by Pettit et al. [25] is combretastatin A-4 (CA-4, 1a, Fig. 1),
which presents remarkable biological activity that is manifested in strong inhibition of tubulin
polymerization, potent in vitro cytotoxicity against a variety of human cancer cell lines,
including multidrug resistant (MDR) cell lines overexpressing P-glycoprotein (Pgp), and in vivo
efficacy as a vascular-disrupting agent (VDA) and antiangiogenic agent [25–29]. Its water-
soluble prodrug, CA-4 disodium phosphate (CA-4P, 1b, Fig. 1), has shown promising results in
numerous clinical trials in both monotherapy and in combination with various anticancer agents,
exhibiting high efficacy in the treatment of anaplastic thyroid cancer and ovarian cancer [30,31].
H CO
3
H
H
3
3
CO
CO
A
A
B
NHAc
H CO
3
B
OCH3
H
3
CO
C
R
O
OCH3
OCH3
R=OH, combretastatin A-4 (CA-4), 1a
R=OPO Na , CA-4P, 1b
Colchicine (2)
3
2
OH
O
O
O
E
A
O
B
C
D
O
H
3
CO
OAc
D
C
O
H CO
3
B
E
OCH3
O
H CO
OCH
3
A
3
O
OCH3
Podophyllotoxin (3)
Steganacin (4)
Fig. 1. Microtubule targeting agents with affinity for colchicine-binding site.
4

ACCEPTED MANUSCRIPT
Promising results in anticancer therapy with CA-4’s unique and multidirectional activity as
well as its relatively simple structure as a lead make extensive structural modification that will
improve its natural anticancer properties possible. Meanwhile, numerous structure-activity
relationship (SAR) studies of CA-4 have been conducted leading to the identification of
functional groups and their positions that are important to ensure optimal interactions with the
colchicine binding site and efficacious antimitotic activity [29,32–37]. In general, these
modifications pertained to the A-ring, B-ring, and/or ethylene bridge. They showed the
importance of the 3,4,5-timethoxy substitution pattern on the A-ring, 4-methoxy substituted B-
ring, and the cis-configured double bond, which are fundamental for the inhibitory activity of
tubulin polymerization.
Bioisosterism is a widely used strategy for the rational design of new drugs including
anticancer agents, particularly for designing agents with optimal pharmacological properties
[38]. The compounds of the studied series of CA-4 derivatives possess an atom of bivalent
oxygen replaced by sulfur. Our previous studies on methylthio-trans-stilbenes as
chemopreventive agents showed the high affinity of selected thio derivatives to human
recombinant CYPs, particulary CYP1A1 and CYP1B1 [39–42]. Other authors showed that the
introduction of a less electronegative sulfur atom instead of the oxygen atom led to a reduction of
toxicity in HEK 293 cells (human embryonic kidney cell line) and thus could improve selectivity
for cancer cell lines [43]. Further modification of the CA-4 molecule, as taken into account in the
design of our series, was an improvement of the stability of the designed CA-4 analogues
because the ethylene bridge in CA-4 is able to undergo cis-trans isomerization, which causes
complete loss of cytotoxicity. This was possible by replacing the olefinic double bond with five-
member heterocyclic rings such as oxazole and N-methylimidazole. Several groups of
5

ACCEPTED MANUSCRIPT
researchers have reported that this type of structural modification allows to avoid the stability
problem and improves anticancer activity [44–54]. Such replacement allows to retain the correct
geometric orientation of the phenyl rings of the CA-4 derivatives by placing them at an
appropriate distance for optimal interaction with the colchicine-binding domain on tubulin.
In this paper we present the design, synthesis, and biological evaluation of a series of CA-4
thio derivatives with different molecular cores. The compounds were designed using a parallel
virtual screening protocol including the generation of a virtual combinatorial library (VCL)
based on an elaborated synthesis protocol of CA-4 analogues, 3-dimensional pharmacophore
screening, two QSAR filters, docking to the colchicine binding site of tubulin, and a final
ranking strategy. The selected derivatives were synthesized and evaluated for their effect on
tubulin polymerization in vitro. Moreover, their antiproliferative activity was assessed in a panel
of six human cancer (A431, HeLa, MCF7, MDA-MB-231, A549, SKOV3) and two non-cancer
cell lines (HaCaT, CCD39Lu). The most potent compounds in the series 23e and 23i were
further studied for cell cycle effects, apoptosis, and on a microtubule network.
2
. Results and discussion
2
.1. Virtual screening
A protocol including combinatorial library generation and screening was developed and
tested to support the design of CA-4 thio analogues (Fig. 2). Based on an elaborated synthetic
protocol for sulfur analogues of CA-4 (Scheme 2-6), different substituted aromatic aldehydes
(BB1) and phenylacetic acids (BB2) were selected from an in-house library and used to generate
the VCL by iterative combination of selected reagents. The obtained VCL (1,159 cmpds) was
processed by the parallel virtual screening (VS) protocol including a 3-dimensional
pharmacophore filter, two QSAR models, and the post-docking scoring method (for details, see
6

ACCEPTED MANUSCRIPT
Supplementary Information). Based on the VS scores, 17 compounds (approx. 2% of VCL) that
had different classification scores (from the worst to the best ones, i.e. 1-5) were selected and
synthesized.
Fig. 2. Illustration of VCL-VS protocol applied for design CA-4 analogues.
To evaluate the real efficiency of the methodology used here, we assumed that the compound
inhibited tubulin polymerization when its IC was lower than 10 ꢀM. Based on this assumption
5
0
and the predicted VS scores, the ROC curve with VS score cut-offs was plotted (Fig. 3). It
showed very good predictive power of the applied VS protocol (AUROC and BEDROC were
0
.85 and 1.00, respectively).
Fig. 3. The ROC curve showing real performance of virtual screening of VCL library.
Interestingly, when the VS score cut-off was greater or equal to 4 the sensitivity (true
positive rate) was 0.43 and precision was 1.00. However, a VS score cut-off greater or equal to 3
7

ACCEPTED MANUSCRIPT
is an optimal choice for virtual screening subjects because both sensitivity and precision are high
(0.86 and 0.67, respectively).
2
.2. Chemistry
CA-4 (1a) was prepared using the two-step procedure described by Gaukroger and co-
workers [34] using Perkin-type condensation between 3,4,5-trimethoxyphenylacetic acid (5) and
isovanillin (6) and subsequent decarboxylation of the obtained
Scheme 1).
α
-phenylcinnamic acid (7)
(
COOH
CHO
H
3
CO
H
H
3
3
CO
CO
H CO
3
a
b
COOH
+
H CO
3
H CO
3
OH
OCH
3
OCH
3
OH
OH
OCH3
OCH3
OCH3
OCH
3
5
6
7
CA-4, 1a
Scheme 1. Synthesis of CA-4. Reagents and conditions: (a) Ac O, Et N, 4 h, 110 ºC, 44%; (b) quinoline,
2
3
Cu, 3 h, 200 ºC, 40%.
Preparation of the designed CA-4 derivatives differing in four molecular cores, namely
phenylcinnamic acids (core 1, 14a-b), ( )-stilbenes (core 2, 15a-b), 4,5-disubstituted oxazoles
core 3, 23a-j), and 4,5-disubstituted N-methylimidazoles (core 4, 24a-f), as cis-restricted
α-
Z
(
analogues is illustrated in Schemes 2-6 following general procedures as detailed below and in the
Experimental Section. Bromobenzaldehyde (10) [55] was prepared from commercially available
vanillin (
with methyl iodide. Methylthiobenzaldehydes 6d [57], 8d [58], 9d
obtained in a multistep reaction starting from appropriately substituted hydroxybenzaldehydes
8
) by its bromination and subsequent methylation of the obtained bromovanillin (
9) [56]
,
11d [57] and 12d [59] were
6
,
8
9
-9 and 11-12 (Scheme 2). In the first step, the obtained O-aryl thiocarbamates 6a [60], 8a [58],
a,
11a [61] and 12a [59] were converted using a Newman-Kwart rearrangement to the
8

ACCEPTED MANUSCRIPT
corresponding S-aryl thiocarbamates 6b, 8b [58], 9b, 11b [61] and 12b [59] which were then
readily hydrolyzed using methanolic potassium hydroxide to thiophenols, and in a one-pot
reaction converted to final products by subsequent methylation with methyl iodide or dimethyl
sulfate. In the preparation of benzaldehydes 9d and 12d, an additional step of protecting the
carbonyl group in S-aryl thiocarbamates 9b and 12b was required by the formation of cyclic
acetals, 1,3-dioxolanes 9c and 12c. Their hydrolysis, methylation, and deprotection were
conducted in a one-pot reaction which allowed to obtain methylthiobenzaldehydes 9d and 12d in
good yields.
CHO
CHO
CHO
a
b
H CO
3
H CO
Br
H
3
CO
Br
3
OH
OH
OCH
3
8
9
10
S
O
c
d
e, f
Ar
1
OH
Ar1 SCH3
Ar O
N(CH
3
)
2
Ar
1
S
N(CH )
3 2
1
6, 8-9, 11-12
6a, 8a-9a, 11a-12a
6b, 8b-9b, 11b-12b
6d, 8d, 11d
CHO
CHO
CHO
g
OCH
3
H CO
Br
3
OCH3
O
O
h, i, j
Ar1 SCH3
6-6d
8-8d
9-9d
Ar1=
9
d, 12d
CHO
CHO
H CO
R
3
SCON(CH 3)2
H CO
OCH3
OCH3
3
9
1
c
R=Br
2c R=OCH3
11-11d
12-12d
Scheme 2. Synthesis of bromo- (9, 10) and methylthiobenzaldehydes (6d, 8d-9d, 11d-12d). Reagents and
conditions: (a) Br , MeOH, 1 h, 0-25 ºC, 99%; (b) methyl iodide, K CO , DMF, 24 h, rt, 88%; (c) H O,
2
2
3
2
KOH, N,N-dimethylthiocarbamoyl chloride, THF, 1-1.5 h, 0-25 ºC, 63-94%; (d) diphenyl ether, 15-120
min, 240 ºC, 54-94%; (e) KOH (10% in MeOH), 2-7 h, 80 ºC; (f) methyl iodide, 24 h, 25 ºC, 67-98% (e
and f); (g) ethylene glycol, p-toluenesulfonic acid monohydrate, toluene, 2.5 h, 145 ºC, 77-94%; (h) KOH
9

ACCEPTED MANUSCRIPT
(
10% in MeOH), 3-7 h, 80 ºC; (i) Me
2 4
SO , 1 h, 25-55ºC; (j) 5% HCl (pH=2-3), 2h, 55ºC, 67-98% (h, i
and j).
The synthesis of
α-phenylcinnamic acids 14a-b [62,63] possessing core 1 and (Z)-stilbenes
1
5a-b [62,63] with core 2 (Scheme 3) was conducted in a manner similar to the preparation of
1
CA-4. The geometries of the (
Z)-stilbenes were confirmed by their characteristic H NMR
coupling constants for olefinic protons of ca. 12–12.1 Hz.
COOH
CHO
H
3
CO
H
H
3
3
CO
CO
H CO
3
R1
R2
COOH
a
R1
R2
b
R1
+
H CO
3
H CO
3
OCH3
OCH3
R
2
OCH3
core 1
core 2
5
11d R
1
=SCH
3
, R
2
=OCH
3
14a R =SCH , R =OCH
3
1
3
2
15a R =SCH , R =OCH
1 3 2 3
13
R =SCH , R =H
14b R =SCH , R =H
1 3 2
15b R =SCH , R =H
1
3
2
1
3
2
Scheme 3. Synthesis of
α-phenylcinnamic acids (14a-b) and (Z)-stilbenes (15a-b). Reagents and
conditions: (a) Ac O, Et N, 4 h, 115 ºC, 34-40%; (b) quinoline, Cu, 2.5 h, 200 ºC, 37-39%.
2
3
The new oxazole-bridged CA-4 analogues 23a-j [62,63] (core 3) (Scheme 5) and N-
methylimidazole-bridged analogues 24a-f [62,63] (core 4) (Scheme 6) were prepared using the
Van Leusen multicomponent reaction between corresponding benzaldehydes and
p-
toluenesulfonylmethyl isocyanides (TosMICs) 20-22 [44,64] whose synthesis is shown in
Scheme 4.
O
-
C
+
H
NH
N
O
S
O
S
CHO
R1
R2
O
R1
R2
O
a
b
R1
R3
R3
R3
R2
CH3
17 R =R =OCH , R =Br
CH3
20 R =R =OCH , R =Br
1
1
1
0 R =R =OCH , R =Br
1
2
3
3
1
2
3
3
1
2
3
3
2d R =R =OCH , R =SCH
18 R =R =OCH , R =SCH
21 R =R =OCH , R =SCH
1 3 3 2 3
1
3
3
2
3
1
3
3
2
3
6
R =R =R =OCH
19 R =R =R =OCH
22 R =R =R =OCH
1 2 3 3
1
2
3
3
1
2
3
3
10

ACCEPTED MANUSCRIPT
2
Scheme 4. Synthesis of TosMIC derivatives (20-22). Reagents and conditions: (a) HCONH , p-
toluenesulfinic acid, CSA, 20 h, 60 ºC, 57-72%; (b) POCl , Et N, DME, 2-3 h, -5 ºC, 70-75%
3
3
The appropriate substituted benzaldehydes obtained earlier, 10 and 12d, or commercially
available 16, were reacted with freshly prepared -toluenesulfinic acid and formamide, catalyzed
p
by 10-camphorsulfonic acid (CSA) to give tosylmethyl formamides 17-19 [44,64]. Dehydration
with POCl afforded the desired TosMICs 20-22. Preparation of N-methylimidazole-bridged
3
analogues (Scheme 6) involved the first generation of aldimines resulting from the reaction of
appropriate benzaldehydes and methylamine and then their condensation with TosMICs in the
presence of anhydrous K CO
2
3.
-
C
+
N
N
O
S
O
R1
CHO
R1
O
R4
R5
R4
R5
a
R2
+
R2
R3
R7
R3
R7
R6
R6
CH3
core 3
6a-d, 8d, 9d, 13
20-22
23a R =R =R =OCH , R =SCH , R5-7=H
1 2 3 3 4 3
2
2
2
2
2
2
2
2
2
3b R =R =R =R =OCH , R =H, R =SC(O)N(CH )
b
b
1 2 3 6 3 4,7 5 3 2
3c R =R =R =R =OCH , R =H, R =SH
1
2
3
6
3
4,7
5
3d R =R =R =OCH , R =SCH , R =H, R =OC(S)N(CH )
3 2
1
3
6
3
2
3
4,7
5
3e R =R =R =OCH , R =SCH , R =H, R =OH
1
3
6
3
2
3
4,7
5
3f R =R =R =R =OCH , R =H, R =SCH
3
1
2
3
5
3
4,7
6
3g R =R =R =R =OCH , R =H, R =Br, R =SCH
3
1
2
3
7
3
4
5
6
3h R =R =R =OCH , R =Br, R =H, R =SCH
1
2
5
3
3
4,7
6
3
3i R =R =R =OCH , R =Br, R =H, R =SCH
1
2
6
3
3
4,7
5
3
3j R =R =R =R =OCH , R =Br, R4,7=H
1
2
5
6
3
3
Scheme 5. Synthesis of oxazole-bridged analogues (23a-j). Reagents and conditions: (a) K CO ,
2
3
DME/MeOH, 2h, reflux, 36-62%; (b) KOH (10% in MeOH), 2 h, 80 ºC, 55-62%.
11

ACCEPTED MANUSCRIPT
-
C
CH
3
+
N
N
N
O
S
N
R
R
1
2
CH
3
CHO
R
R
O
1
2
R
R
4
5
R4
a
b
R
R
4
5
H
3
C
NH
2
+
+
R
7
R7
R5
R
R
3
3
R
7
R
6
R6
R
6
CH
3
core 4
6
a, 6d, 8d, 9d, 13
20-22
24a R
4b R =R =R =OCH , R =SCH , R =H, R =OC(S)N(CH )
3 2
1 2 3 3 4 3
=R =R =OCH , R =SCH , R5-7=H
2
c
1
3
6
3
2
3
4,7
5
2
2
2
2
4c R =R =R =OCH , R =SCH , R =H, R =OH
1 3 6 3 2 3 4,7 5
4d R =R =R =R =OCH , R =H, R =Br, R =SCH
3
1
2
3
7
3
4
5
6
4e R =R =R =OCH , R =Br, R =H, R =SCH
3
1
2
5
3
3
4,7
6
4f R =R =R =OCH , R =Br, R =H, R =SCH
3
1
2
6
3
3
4,7
5
Scheme 6. Synthesis of N-methylimidazole-bridged analogues (24a-f). Reagents and conditions: (a)
AcOH, EtOH, 2h, reflux; (b) K
h, 80 ºC, 33%.
2 3
CO , DME, 3h, reflux, 27-92% (a and b); (c) KOH (10% in MeOH), 5.5
Preparation of heterocycle-based derivatives 23c, 23e, and 24c possessing the mercapto or
hydroxy group involved synthesis of their S-aryl thiocarbamate or O-aryl thiocarbamate
analogues followed by basic hydrolysis of the carbamate groups.
2
2
.3. Biological evaluation
.3.1. In vitro cytotoxic activities
Table 1 summarizes the cytotoxic effects of the obtained CA-4 analogues with each
molecular core, namely
α
-phenylcinnamic acids (14a-b), (Z)-stilbenes (15a-b), 4,5-disubstituted
oxazoles (23a 23c 23e-j), and 4,5-disubstituted N-methylimidazoles (24a,
,
,
24c-f), against a
panel of six human cancer cell lines (A431, HeLa, MCF7, MDA-MB-231, A549, and SKOV)
and two human non-cancer cell lines (HaCaT and CCD39Lu), using CA-4 as the reference
compound.
Table 1. Cytotoxic activities of 14a-b, 15a-b, 23a, 23c, 23e-j, 24a, 24c-f and CA-4 against human cancer
and non-cancer cell lines
12

ACCEPTED MANUSCRIPT
N
COOH
14a R =SCH , R =OCH
3
1
3
2
23a R
=R =R =OCH , R =SCH , R5-7=H
1 2 3 3 4 3
O
1
4b R =SCH , R =H
R
1
23c R =R =R =R =OCH , R =H, R =SH
1
3
2
1
2
3
6
3
4,7
5
H CO
3
2
3e R =R =R =OCH , R =SCH , R =H, R =OH
A
1 3 6 3 2 3 5
4,7
=H, R =SCH3
A
R4
R5
23f R =R =R =R =OCH , R
R1
R2
1
2
3
5
3
4,7
6
R2
H CO
3
23g R =R =R =R =OCH , R =H, R =Br, R =SCH
1 2 3 7 3 4 5 6 3
B
B
R3
23h R =R =R =OCH , R =Br, R
6
=H, R =SCH3
OCH3
1
2
5
3
3
4,7
R7
23i R =R =R =OCH , R =Br, R =H, R =SCH3
1
2
6
3
3
4,7
5
R6
23j
R =R =R =R =OCH , R =Br, R4,7=H
1
2
5
6
3
3
α-phenylcinnamic acids
core 1)
4
,5-disubstituted oxazoles
(
(
core 3)
N
2
4a R =R =R =OCH , R =SCH , R5-7=H
1
5a R =SCH , R =OCH
3
1 2 3 3 4 3
1
3
2
N
R1
CH
3
24c R =R =R =OCH , R =SCH , R =H, R =OH
1 3 6 3 2 3 4,7 5
15b R =SCH , R =H
H CO
3
1
3
2
2
4d R =R =R =R =OCH , R =H, R =Br, R =SCH
A
1 2 3 7 3 4 5 6 3
A
R4
24e R =R =R =OCH , R =Br, R
=H, R =SCH3
1 2 5 3 3 4,7 6
R1
R2
R2
H CO
3
24f R
1
=R
2
=R
6
=OCH
3
, R
3
=Br, R4,7=H, R =SCH
5 3
B
B
R
3
OCH
3
R7
R5
R6
(
Z)-stilbenes
core 2)
4,5-disubstituted N-methyl-
imidazoles
(
(
core 4)
a
IC50 (µM)
Compd
A431
HeLa
MCF7
MDA-MB-231
A549
SKOV3
HaCaT
CCD39Lu
1
4a
4b
5a
5b
3a
>20
>20
>20
>20
0.95±0.33
1.84±0.36
0.65±0.19
2.35±1.04
1.43±0.37
2.11±0.57
0.45±0.14
0.60±0.10
1.42±0.23
2.48±0.64
2.78±0.77
6.10±1.05
2.29±0.46
1.63±0.27
2.51±0.87
>20
>20
>20
>20
>20
>20
>20
>20
>20
1
>20
1
3.61±2.44 6.18±0.46
8.22±2.66 11.68±2.54
11.73±0.85
14.20±0.54
>20
>20
14.47±1.25
14.08±0.61
>20
9.56±3.03
12.82±1.83
>20
>20
1
>20
>20
2
>20
7.33±0.91
>20
>20
2
2
3c
3e
1.22±0.82 4.44±0.97
0.25±0.20 0.009±0.002
0.43±0.29 0.645±0.03
2.87±0.34
0.71±0.16
0.85±0.13
17.87±0.90
0.76±0.10
0.92±0.14
7.83±1.18
>20
>20
3.18±0.26
0.25±0.12
1.48±0.16
>20
1.28±0.15
0.32±0.09
0.43±0.13
16.74±2.64
0.26±0.10
0.52±0.08
10.09±1.96
>20
>20
>20
>20
2
3f
3g
3h
>20
>20
2
>20
>20
>20
>20
2
0.41±0.10 0.44±0.04
0.43±0.11 0.63±0.04
12.32±0.25 0.07±0.04
>20
1.03±0.12
1.16±0.13
0.26±0.18
>20
>20
2
3i
>20
>20
2
3j
0.50±0.18
>20
>20
2
4a
4c
4d
>20
>20
>20
2
0.43±0.13 0.39±0.02
16.57±3.71 8.18±2.28
0.39±0.28
>20
>20
1.72±0.14
>20
0.50±0.09
>20
>20
2
>20
6.13±0.75
13

ACCEPTED MANUSCRIPT
2
4e
13.64±3.23
>20
>20
>20
2.20±0.57
5.24±1.11
0.17±0.04
15.45±0.37
>20
>20
>20
>20
>20
>20
0.38±0.09
>20
>20
19.26±2.14
>20
2
4f
CA-4
0.25±0.08 0.11±0.06
0.56±0.08
0.19±0.08
0.009±0.002
a
IC – compound concentration required to inhibit cell proliferation by 50%. Data are expressed as the
50
mean ± SD from dose-response curves of three independent experiments.
The results listed in Table 1 indicate that cytotoxicity toward cancer cells varied between
each molecular core. Compounds 14a-b as analogues possessing the -phenylcinnamic acid core
core 1) were generally inactive against the cancer cell lines (IC50 > 20 M), except for MCF7
from estrogen-dependent breast cancer adenocarcinoma, toward which they displayed relatively
high antiproliferative activity (IC50 values of 0.95-1.84 M).
α
(
µ
µ
In turn, compounds 15a-b containing the (Z)-stilbene scaffold (core 2) with the same
substituents in the phenyl rings displayed slightly enhanced activity toward four cancer cell lines
(
A431, HeLa, MDA-MB-231, and SKOV3) with IC values of 3.61-14.47 µM, while activity
50
against MCF7 cells was retained.
The oxazole-bridged CA-4 analogues (core 3) generally displayed the best antiproliferative
activity from all of the tested compounds toward the cancer cell lines except for A549, toward
which most of the compounds were inactive. The number and position of the substituents in the
phenyl B-ring had a major influence on antiproliferative activity. Compounds with substituents
at 3,4-position of the phenyl B-ring showed much higher activity than compound 23a with the -
SCH group at the orto-position or compound 23g with substituents at 3,4,5-position in the
3
phenyl B-ring. Replacement of the
3f) with an electron withdrawing bromine atom (compound 23h) maintained antiproliferative
activity.
m-OCH group in the trimethoxyphenyl A-ring (compound
3
2
14

ACCEPTED MANUSCRIPT
The most active compound, 23e, in this series and of all the tested compounds was over 10-
fold more potent against HeLa cells than the reference compound CA-4 with an IC value of
5
0
0
.009
µ
M vs 0.11
µ
M, respectively. Interestingly, bioisosteric replacement of the oxygen atom
–SCH group in compound 23i instead of the –OCH3
with a sulfur atom (introduction of the
m
m
3
group in compound 23j) led to an increase in antiproliferative activity toward the A431 (28.7-
fold), MCF7 (2.2-fold), and MDA-MB-231 (8.5-fold) cell lines. On the other hand, only
compound 23j was active against the highly drug-resistant A549 lung adenocarcinoma with an
IC50 value of 0.5
N-methylimidazole-bridged CA-4 analogues (24a
less active than their corresponding oxazole analogues (23a
results of 23a vs 24a 23h vs 24e, and 23i vs 24f, thus suggesting that replacement of the oxazole
µM.
,
24e, and 24f) (core 4) were significantly
23h, and 23i) when compared to the
,
,
with the N-methylimidazole moiety was the major reason for the loss of activity in these
compounds. Compound 24c as an imidazole-bridged analogue of 23e was only active in this
series.
Because the toxicity of antitumor drugs toward normal tissues is a very important issue in
chemotherapy, we evaluated the cytotoxic effects of these derivatives on normal human cells.
For this purpose, all compounds were tested in vitro in human immortalized keratinocytes
HaCaT and lung fibroblasts CCD39Lu. Based on the results, most of the compounds, with the
exception of 24d, showed slight cytotoxicity on fibroblast cells (IC > 20
µ
M) in contrast to the
M). However, some of the tested derivatives (23c
23i and 24c) were cytotoxic against immortalized human keratinocytes HaCaT
IC50 values of 0.32-1.28 M verus 0.19 M for CA-4).
5
0
reference compound CA-4 (IC of 0.009
µ
,
5
0
23e, 23f, 23h,
(
µ
µ
15

ACCEPTED MANUSCRIPT
Finally, two of the most active compounds (23e and 23i) and six of the most sensitive cell
lines (A431, HaCaT, HeLa, MCF-7, MDA-MB-231, and SKOV-3) were selected for further
studies such as assessment of the cell cycle, tubulin staining (effect on microtubule dynamics),
and apoptosis induction.
2
.3.2. Inhibition of tubulin polymerization
Compounds 14a-b 15a-b 23a 23c 23e-j, 24a, and 24c-f were investigated for their effect
,
,
,
,
on tubulin polymerization with the use of turbidimetric assay in vitro. The N-methylimidazole-
bridged analogues, unlike oxazole-bridged analogues, were poor tubulin polymerization
inhibitors (Table 2). In the series of oxazole-bridged analogues, compounds 23a
, 23e, and 23i
efficiently inhibited tubulin polymerization with an IC of 0.86, 1.05, and 0.85 M, respectively
µ
5
0
(Table 2). CA-4 was used as a reference compound.
Table 2. Inhibition of tubulin polymerization and VS score for compounds 14a-b, 15a-b, 23a, 23c, 23e-j,
4a, 24c-f and CA-4
2
a
b
Compd
4a
IC (µM)
VS score
50
1
>10
>10
>10
>10
1
2
3
2
3
4
5
3
14b
15a
15b
23a
23c
23e
0.86 (0.54-1.37)
2.97 (2.36-3.73).
1.05 (0.65-1.69)
1.62 (0.94-2.80)
23f
16

ACCEPTED MANUSCRIPT
2
3g
>10
1.54 (1.33-1.79)
0.85 (0.54-1.31)
4.80 (3.69-7.26)
>10
2
3
2
3
3
4
1
2
3
5
2
3h
2
3i
2
3j
2
4a
4c
2
6.67 (6.37-6.98)
>10
2
4d
4e
4f
CA-4
2
>10
2
>10
2.72 (2.20-3.37)
a
Tubulin polymerization was monitored spectrophotometrically at 340 nm for 1 h at 37˚C. The
experiments were performed twice in duplicate. In parentheses 95% confidence intervals determined with
the use of GraphPad Prism 5 are presented.
b
Virtual Screening (VS) score was obtained according to the ranking scheme described in Experimental
Section.
Compound 23j as an oxygen analogue of 23i weakly inhibited tubulin polymerization with
IC50 equal to 4.80 µM. It should be noted that this outcome is in line with predictions obtained in
the post-docking scoring strategy combining Tanimoto and SIFt (for details, see Supplementary
Information), where the sulfur analogue was scored higher than the oxygen analogue.
2
.3.3. Analysis of cell cycle and tubulin staining
In this experiment, the effect of the tested compounds on the cell cycle was assayed in six
selected cancer cell lines; camptothecin at a concentration of 50 nM was used as a positive
control (Fig. 4).
17

ACCEPTED MANUSCRIPT
Fig. 4. Effect of 23e, 23i and CA-4 on cell cycle distribution in A431, HaCaT, HeLa, MCF-7, MDA-MB-
2
31 and SKOV-3 cells. Cell cycle was monitored by flow cytometric analysis at 24 h after cells were
treated with tested compounds. Data are expressed as mean ± SD of three independent experiments: (*) p
0.05, (**) p < 0.01. Statistical significance between groups was assessed by Dunnett’s Multiple
<
Comparison Test.
Although all three of the tested compounds were shown to be able to inhibit polymerization
of tubulin (Table 2), their impact on the cell cycle was significantly cell line-dependent.
18

ACCEPTED MANUSCRIPT
Interestingly, very similar changes in cell cycle distribution were observed in the HeLa and
SKOV-3 cell lines. Compound 23e induced a statistically significant concentration-independent
block of SKOV-3 cells in G2/M, while a similarly massive G2/M block was observed after
incubation with CA-4. In contrast, after incubation with compound 23i in HeLa and SKOV-3
cells, the concentration-dependent block of the G2/M phase was shown.
A similar trend was observed in A431, HaCaT, MCF-7, and MA-MB-231 cells treated with
compounds 23e, 23i, and CA-4. The tested compounds were also investigated for their inhibition
of tubulin polymerization using immunostaining (Fig. 5).
19

ACCEPTED MANUSCRIPT
Fig. 5. Effect of 23e, 23i and CA-4 on tubulin polymerization and microtubules and nuclear condensation
in A431, HeLa, HaCaT, MDA-MB-231, MCF-7 and SKOV-3 cells. Cells were treated with 23e, 23i and
CA-4 at concentration of 1 µM for 24h. The control (intact) cells are presented in the first column on the
left. The morphology of microtubules was visualized by staining with anti-alpha-tubulin antibodies
connected with FITC (green), while the morphology of nuclei condensation was shown by staining with
Hoechst 33258 (blue).
20

ACCEPTED MANUSCRIPT
It is worth noting that besides disturbances in tubulin polymerization as shown in all of the tested
cells, condensation of chromatin (stained with Hoechst 33358 dye) which is also typical for
apoptosis, could be observed in the nuclei of cells incubated with 23i (Fig. 5). Perturbations in
the cytoskeleton structure as well as micronuclei formation corresponded with the results of
ELISA and flow cytometry studies, where clear differences between effects exerted on cells by
2
3e, CA-4, and, to a lesser extent, by 23i could also be observed.
.3.4. Induction of apoptosis
2
Induction of apoptosis in cells incubated with compounds 23e, 23i, and CA-4 was analyzed
using two different methods. The ELISA assay measures the presence of markers for apoptotic
cells, which was based on the quantitative detection of histone-associated DNA fragments in
mono- and oligonucleosomes, while staining with Annexin V allows to detect phosphatidylserine
externalization. Translocation of phosphatidylserine from the inner to the outer leaflet of the
membrane is observed in intermediate stages of apoptosis and may be shown using Annexin V-
FITC staining followed by flow cytometry. For cells incubated with 23e, 23i, and CA-4, a clear
concentration-dependent statistically significant increase of Annexin V positive cells was
observed (Fig. 6).
Interestingly, in cells incubated with 23e, the highest level of histone-associated DNA
fragments in mono- and oligonucleosomes was observed only at the lowest concentration of 23e
and CA-4, while changes observed in cells incubated with 23i may be described as
concentration-dependent (Fig. 7).
21

ACCEPTED MANUSCRIPT
Fig. 6. CA-4 thio derivatives 23e and 23i increase apoptosis in A431, HaCaT, HeLa, MCF-7, MDA-MB-
2
31 and SKOV-3 cells. Apoptosis was detected after 24 h of incubation with tested compounds by
Annexin V-FITC/PI double staining by flow cytometry. Data are expressed as mean ± SD of three
22

ACCEPTED MANUSCRIPT
independent experiments. Statistical significance between groups was assessed by Dunnett’s Multiple
Comparison Test. All treated groups were significantly different from control at p < 0.05.
PLUS
Fig. 7. Apoptosis detection using Cell Death Detection ELISA
followed by 24 h incubation of A431,
HaCaT, HeLa, MCF-7, MDA-MB-231 and SKOV-3 cells with 23e, 23i and CA-4. Data are expressed as
23

ACCEPTED MANUSCRIPT
mean ± SD of three independent experiments. Statistical significance is indicated with asterisks: ***p <
0
.001, **p < 0.01 and *p < 0.05 indicate a significant difference from the control.
These results may suggest significant differences in the cellular uptake, metabolism, and
cellular efflux of these cells as well as coexistence of different cellular signaling pathways
stimulated by these compounds, although subsequent tests are necessary in order to fully explain
these mechanisms.
2
.3.5. Molecular modeling
Docking studies revealed that the binding modes of the most active compounds (23e and 23i)
are coherent with co-crystallized DAMA-colchicine (Fig. 8). The trisubstituted phenyl A-ring of
3e and 23i is buried in the β-subunit occupying the same position as the corresponding ring of
2
colchicine.
Fig. 8. The binding mode of A) 23e (green) and B) 23i (yellow) with DAMA-colchicine (magenta) in the
β-tubulin binding site (PDB code 1SA0). Hydrogen bonding was depicted as orange dotted lines.
The thiol group of Cys241β forms a hydrogen bond with the sulfur or oxygen atom of the
methylthio or methoxy group in para-position in 23e and 23i, respectively. In the most active
24

ACCEPTED MANUSCRIPT
compound 23e, the -SCH
3
group (ring A) is involved in hydrophobic interaction with the side
chain of Leu255β, Ala316β, and Ile378β. The bromine atom in 23i is placed in a position that
allows hydrophobic interactions with Cys241β, Ala250β, and Leu255β. Moreover, the -SCH3
substituent (23i, ring B) remains in close contact with Ala316β, Leu255β, and Met259β, thus
making non-bonded S⋅⋅⋅S interaction with the sulfur atom of Met259β possible. It should be
noted that oxazole derivatives (23e and 23i) showed a similar binding mode with other cis-
restricted analogues of CA-4 containing five-member aromatic heterocyclic rings, such as
thiazoles [47], triazoles [46], and terazoles [48] with the trimethoxyphenyl ring placed in
proximity of Cys241β.
3
. Conlusions
We designed and synthesized a series of CA-4 thio derivatives containing different molecular
cores, namely
α-phenylcinnamic acids (core 1), (Z)-stilbenes (core 2), 4,5-disubstituted oxazoles
(core 3), and 4,5-disubstituted N-methylimidazoles (core 4) as cis-restricted analogues.
Compounds for synthesis were selected with the use of a parallel virtual screening protocol
which showed a good predictive power (AUROC and BEDROC values were 0.85 and 1.00,
respectively).
Biological evaluation of the tested compounds revealed the highest antiproliferative activities
in the group of oxazole-bridged analogues (core 3) with the 3,4,5-trisubstituted phenyl A-ring
and 3,4-disubstituted phenyl B-ring. The compounds 23e, 23f, 23h, 23i, and 24c (N-
methylimidazole-bridged analogue) displayed pronounced cytotoxic activity against the cancer
cell lines: A431, HeLa, and MDA-MB-231. Compound 23e was over 10-fold more potent in
inhibiting HeLa cell proliferation with IC50 values of 0.009 µM than positive control CA-4.
Additionally, the compounds 23a, 23e, and 23i were the most potent inhibitors of tubulin
25

ACCEPTED MANUSCRIPT
polymerization determined in vitro. Compound 23i, a sulfur analogue of 23j, showed 5-fold
higher inhibitory effectiveness on tubulin polymerization compared to 23j. The influence of
exchange of oxygen to sulfur could be a promising subject for further studies.
Moreover, compounds 23e and 23i, which showed both the highest cytotoxic and antitubulin
activities, were further studied in terms of their effect on cell cycle distribution and proapoptotic
activity. The impact of 23e and 23i on cell proliferation was cell line-dependent. A statistically
significant block of the cell cycle at the G2/M phase was observed for compound 23e in A431,
HaCaT, HeLa, MCF-7, MDA-MB-231, SKOV-3 cells, and for compound 23i in HeLa and
SKOV-3 cells. The proapoptotic effect of 23e and 23i in studied cell lines was demonstrated with
ELISA assay and double staining with Annexin V-FITC/PI. The results of our studies indicate
that the CA-4 derivatives 23e and 23i may serve as novel lead compounds in research on more
effective anticancer agents.
4
. Experimental section
4
.1. General methods for chemistry
All reagents and solvents were purchased from Sigma-Aldrich, Acros, Fluka, and POCH and
were used as received. Reactions that involved air or moisture-sensitive reagents were performed
in oven-dried glassware under an inert atmosphere of dry nitrogen with dried solvents, unless
otherwise stated. The progress of all reactions was monitored on Merck precoated silica gel
plates (with fluorescence indicator UV254) and visualized in UV light (λmax 254 or 365 nm).
Melting points were determined in capillary tubes on a Stuart SMP10 micro melting point
1
13
apparatus and are uncorrected. H NMR and C NMR spectra were recorded at the Institute of
Bioorganic Chemistry, Polish Academy of Sciences in Poznań, using Bruker 400 (400 MHz for
1
13
1
13
H and 101 MHz for C), Bruker 500 (500 MHz for H and 126 MHz for C), and Bruker 700
26

ACCEPTED MANUSCRIPT
1
13
(
700 MHz for H and 176 MHz for C) spectrometers with TMS as an internal standard in
CDCl or DMSO-d6. Chemical shifts (δ) are quoted in parts per million (ppm) and are referred to
3
1
13
as a solvent residual peak (CDCl , δ 7.26 ppm for H and δ 77.0 ppm for C NMR; DMSO-d6
3
1
13
δ 2.5 ppm for H and δ 39.5 ppm for C NMR). Coupling δ constants (J) are quoted in Hertz
(Hz) and peaks are listed as singlet (s), doublet (d), triplet (t), multiplet (m), doublet of doublets
1
1
(
dd), triplet of doublets (td), and broad signal (bs). Additional techniques ( H− H COSY, HSQC,
HMBC) were used to assist allocation. LRMS (EI) spectra were recorded on a Bruker
20MS/420GC mass spectrometer and HRMS (EI) spectra were recorded on an Intectra Mass
AMD 402 or 604 mass spectrometer by the Advanced Chemical Equipment and Instrumentation
Facility at the Faculty of Chemistry, Adam Mickiewicz University in Pozna . IR spectra were
recorded on a Bruker FT-IR IFS 66v/s spectrometer by the Advanced Chemical Equipment and
Instrumentation Facility at the Faculty of Chemistry, Adam Mickiewicz University in Pozna or
3
ń
ń
on a Thermo Scientific Nicolet iS5 FT-IR spectrometer at the Institute of Mathematical and
Natural Sciences, Department of Chemistry, State Higher Vocational School in Tarnów. Dry
flash column chromatography was carried out on Merck silica gel 60, particle size 40−63 ꢀm or
1
5-40 ꢀm using EZSafe low-pressure columns. UV−vis spectra were recorded on a Hitachi
UV/vis U-1900 spectrophotometer; λmax (log ε), nm. All final target compounds were
characterized and determined to be at least >95% pure using a Waters ACQUITY UPLC H-class
system equipped with a UV DAD and TQD Waters MS detector with electrospray ionization at
the Department of Medicinal Chemistry, Institute of Pharmacology, Polish Academy of Sciences
in Kraków. LC analysis was performed using a ACQUITY UPLC BEH C18 column (2.1 × 50
mm, 1,7 ꢀm) at a flow rate of 0.3 mL/min (20–100% aqueous CH CN over 3 min, 100% CH CN
3
3
over 0.5 min, and 100–20% aqueous CH CN over 2.5 min).
3
27

ACCEPTED MANUSCRIPT
4
4
.2. General procedure for synthesis of compounds 1a
.2.1. Synthesis of CA-4 (1a
, 14a-b, 15a-b, 23a-j, and 24a-f
)
CA-4 was synthesized according to known procedures [34].
.2.1.1. (E)-3-(3-Hydroxy-4-methoxyphenyl)-2-(3,4,5-trimethoxyphenyl)prop-2-enoic acid (
A mixture of 3-hydroxy-4-methoxybenzaldehyde 6 (1.36 g, 8.94 mmol), 3,4,5-
trimethoxyphenylacetic acid 5 (2.01 g, 8.90 mmol), acetic anhydride (4 mL), and trimethylamine
2 mL) were heated under nitrogen at 110 °C for 4 h. The mixture was cooled and 6 mL of conc.
4
7)
(
HCl and 6 mL of diluted HCl was added. A yellow solid was obtained which was collected by
filtration. It was dissolved in 10% aq NaOH (50 mL) and washed with ethyl acetate (3x5 mL).
The organic layers were separated and the aqueous layer was acidified with conc. HCl to pH 3-4.
The precipitated crude product was collected by filtration and recrystallized from 70% v/v EtOH
with the use of activated carbon to afford 7 as pale yellow crystals. Yield 44% (1.40 g, 3.90
mmol). mp 239-241 °C (lit. [34] 237-239 °C). R (EtOAc/n-hexane, 3:1) 0.18. UV-vis (MeOH)
f
1
λ
max (log ε): 322 (4.64), 235 (4.67). HPLC: 95.71%, tR = 2.18 min. H NMR (400 MHz, DMSO-
d
8
3
6
) δ 8.94 (bs, 1H, COOH), 7.57 (s, 1H, C1a’-H), 6.80 (d, J = 8.5 Hz, 1H, C5’-H), 6.60 (dd, J =
.5, 2.0 Hz, 1H, C6’-H), 6.53 (d, J = 2.0 Hz, 1H, C2’-H), 6.44 (s, 2H, C2-H, C6-H)), 3.73 (s,
1
3
H, C4-OCH ), 3.71 (s, 3H, C4’-OCH ), 3.69 (s, 6H, C3-OCH , C5-OCH ). C NMR (101
3
3
3
3
MHz, DMSO-d ) δ 168.57 (COOH), 153.07 (C3, C5), 148.88 (C3’), 145.85 (C4’), 139.08 (C4),
6
1
36.95 (C1), 132.14 (C1’), 130.33 (C1a), 127.03 (C1a’), 122.92 (C6’), 117.21 (C2’), 111.52
(
C5’), 106.73 (C2, C6), 60.12 (C4-OCH ), 55.94 (C3-OCH , C5-OCH ), 55.46 (C4’-OCH ).
3
3
3
3
-
1
FTIR ATR (KBr) cm : 3313 (bs
O-CH ), 1666 ( C=O), 1584 (
s
C-O), 1237 (as Ar-O-C), 1123 (ν C-O-C), 1023 (ν CAr-O-C), 900 (γ OH…O) 805
ν
O-H), 3073 (
ν
= CAr-H), 2961 (
ν
as CH
3
), 2939 (
ν
CH
3
), 2826
(
ν
ν
ν C =C ), 1503 (ν C =C ), 1455 (δ CH ), 1409 (δ CH ),
Ar Ar Ar Ar as 3 s 3
3
oop
1
265 (
ν
C
28