Development of a novel class of tubulin inhibitor from desmosdumotin B with a hydroxylated bicyclic B-ring

Article abstract of DOI:10.1021/jm501859j

A series of newly synthesized hydroxylated analogues of triethyldesmosdumotin B (TEDB) with a bicyclic B-ring exhibited a significantly different mode of action for affecting microtubule dynamics and spindle formation but had the same antiproliferative ac

Full text of DOI:10.1021/jm501859j

Article
pubs.acs.org/jmc
Development of a Novel Class of Tubulin Inhibitor from
Desmosdumotin B with a Hydroxylated Bicyclic B‑Ring
Kyoko Nakagawa-Goto,*,†,‡ Akifumi Oda,† Ernest Hamel,∥ Emika Ohkoshi,‡ Kuo-Hsiung Lee,‡,§,⊥
and Masuo Goto*,‡
†School of Pharmaceutical Sciences, College of Medical, Pharmaceutical and Health Sciences, Kanazawa University, Kanazawa,
920-1192, Japan
‡Natural Product Research Laboratories, Chemical Biology and Medicinal Chemistry, UNC Eshelman School of Pharmacy, University
of North Carolina, Chapel Hill, North Carolina 27599-7568, United States
§UNC Lineberger Comprehensive Cancer Center, University of North Carolina, Chapel Hill, North Carolina 27599-7295, United
States
∥Screening Technologies Branch, Developmental Therapeutics Program, Division of Cancer Treatment and Diagnosis, Frederick
National Laboratory for Cancer Research, National Cancer Institute, Frederick, Maryland 21702, United States
⊥Chinese Medicine Research and Development Center, China Medical University and Hospital, 2 Yuh-Der Road, Taichung, 40447,
Taiwan
S
* Supporting Information
ABSTRACT: A series of newly synthesized hydroxylated
analogues of triethyldesmosdumotin B (TEDB) with a bicyclic
B-ring exhibited a significantly different mode of action for
affecting microtubule dynamics and spindle formation but had
the same antiproliferative activity spectrum, including activity
against multidrug-resistant tumors. These analogues efficiently
induced cell cycle arrest at prometaphase and caused
formation of immature multipolar spindles. 6′-Hydroxyl
TEDB-TB (8) disrupted bipolar spindle formation but had a
negligible effect on interphase microtubules. On the basis of the predicted binding modes of the new compounds with tubulin
dimer, compound 4 forms three hydrogen bonds (H-bonds) only with α-tubulin at the colchicine site; in contrast, 8 forms H-
bonds with both α- and β-tubulin. We predict that, when a compound/ligand, such as 8, forms H-bonds to both α- and β-
tubulins, spindle formation is disrupted more than the dynamics of interphase microtubules. This result may reflect the well-
known greater dynamicity of spindle microtubules as compared with interphase microtubules.
antimicrotubule agents have been identified and developed
from natural products. Antimicrotubule agents, especially
taxanes and vinca alkaloids, are commonly used in clinical
cancer treatment. Furthermore, many microtubule-targeting
natural products are currently in clinical trials. However,
although antimicrotubule agents are quite useful cancer
chemotherapeutics, one major obstacle remains unsolved:
innate and acquired drug resistance, notably multidrug
resistance (MDR). The best-known gene is MDR-1, encoding
P-glycoprotein (P-gp). Overexpression of the P-gp drug
transporter leads to poor disease prognosis,5−9 and most
clinical antimicrotubule drugs are P-gp substrates. [A survey of
numerous antimicrotubule agents with a single matched pair of
cells, with one of the lines overexpressing P-gp, demonstrated
that PXL, docetaxel, ixabepilone, VIN, vinblastine, vinorelbine,
and the truncated halichondrin B analogue eribulin were P-gp
substrates (Hamel and R. Bai, unpublished data).] A recent
1. INTRODUCTION
Microtubule-targeting agents are recognized as valuable drugs
in the treatment of both solid cancers and hematologic
malignancies.1−3 Antimicrotubule agents disrupt microtubule
dynamics, resulting in disruption of cell cycle progression and,
subsequently, induction of cell death. On the basis of the
mechanisms of action, antimicrotubule agents are normally
categorized into two groups: (i) microtubule disrupters that
inhibit the polymerization of tubulin, such as colchicine,
vincristine (VIN), podophyllotoxin, and combretastatin A-4
(CA-4), and (ii) microtubule stabilizers that prevent the
depolymerization of microtubules, such as paclitaxel (PXL),
epothilones, discodermolide, and laulimalide. Depending on
their binding domain on tubulin, these antimitotic agents are
further classified into (i) colchicine site agents (CSAs), such as
colchicine and CA-4; (ii) vinca site agents, such as VIN and
vinblastine; (iii) taxane site (on the microtubule interior)
agents, such as PXL and epothilones; (iv) laulimalide site (on
the microtubule exterior) agents, such as laulimalide and
peloruside A; and (v) others.4 Importantly, numerous
Received: December 2, 2014
Published: February 19, 2015
© 2015 American Chemical Society
2378
DOI: 10.1021/jm501859j
J. Med. Chem. 2015, 58, 2378−2389

Journal of Medicinal Chemistry
Article
Figure 1. Structures of desmosdumotin B and related bioactive analogues.
Figure 2. Structures of TEDB-TB analogues.
a
Scheme 1. Syntheses of New Analogues of 4 and 5
a
Reagents and conditions: (a) ArCHO, 50% KOH, EtOH, rt; (b) I2 (cat.), DMSO, H2SO4 (cat.), 90−95 °C, 1 h; (c) BBr3, 0 °C to rt.
study indicated that overexpression of class III β-tubulin, as well
as changes in microtubule regulation, are also associated with
MDR to antimicrotubule agents that interact with the taxane or
vinca site.10 However, interestingly, CSAs were generally active
in cells overexpressing class III β-tubulin.11
cells, was >20. This selectivity against MDR lines was greatly
enhanced by triethylation of the A-ring, producing 2 (TEDB, SI
> 220) and 3 (SI = 490).12−14 The TEDB analogues with a
bicyclic aryl B-ring, such as benzo[b]thiophene 4 (TEDB-TB)
and naphthalene 5, displayed a dramatically different bioactivity
profile, showing substantial growth inhibition against multiple
cancer cell lines.13,15 These analogues inhibited tubulin
polymerization, in part through the CS. Importantly, none of
the analogues were P-gp substrates, and thus, they may be
potentially effective against MDR tumors.15
Poor water solubility is another major obstacle for the use of
antimitotic agents, especially the taxanes.2 Generally, the
incorporation of a hydroxyl group into a structure can lead to
increased polarity as well as polar surface area (PSA), which is
connected to oral bioavailability16 and correlated with passive
intestinal absorption.17,18 cLogP has been used as a parameter
of water solubility. In the TEDB-TB series, cLogP values of the
related hydroxyl analogues were reduced by a range of 0.5−0.9
from those of the parent compounds (Supporting Information,
Table S1).19 In a prior modeling study, the structure of TEDB-
Although rapidly growing cells, such as tumors, are sensitive
to antimicrotubule agents, normal cells are also impacted by
disruption of microtubule networks. Because no CSAs are
currently used clinically for cancer treatment, the discovery of
new antimitotic agents targeting the colchicine site (CS) might
be a valuable approach for effective cancer chemotherapy,
especially agents that can overcome the MDR phenotype
without interfering with interphase microtubule dynamics.
Desmosdumotin B (DesB, 1), a flavonoid with an unusual
nonaromatic trimethylated A-ring (Figure 1), exhibited unique
bioactivity against P-gp overexpressing MDR tumor cells.
Briefly, 1 displayed selective antiproliferative activity against
tumor cell lines overexpressing P-gp, while it was nontoxic
against non-MDR tumor cell lines. Its selective index (SI), the
IC50 against non-MDR cells divided by the IC50 against MDR
2379
DOI: 10.1021/jm501859j
J. Med. Chem. 2015, 58, 2378−2389

Journal of Medicinal Chemistry
Article
TB partially matched with the common pharmacophores
conserved in antimicrotubule CSAs.15 This same study also
suggested that a hydrogen acceptor, such as a hydroxyl group,
on the B-ring in TEDB-TB could contribute to an additional
common CS-related pharmacophore. Inspired by these
analyses, we rationally designed and synthesized various
hydroxyl analogues of 4 and 5 to improve water solubility, as
well as to study structure−activity relationships (SARs). A
hydroxyl group also allows various functional groups to be
introduced into the compound in future modification studies.
All synthesized analogues were evaluated for antiproliferative
activity against several cancer cell lines, including an MDR
tumor cell line. Active analogues were further assessed for
potential effects against tubulin assembly in a cell-free system
and cell cycle progression in tumor cells. We also performed
molecular modeling studies with the analogues bound to
tubulin, and these demonstrated a novel mode of H-bond
formation in the CS.
from 32, 4- and 6-methoxy derivatives (36 and 37) from 33,
and the 7-methoxy derivative (38) from 34.20 The C-3 methyl
group of 35−38 was converted by radical bromination to a
bromomethyl group, which was then formylated by using a
Sommelet reaction to obtain the desired 3-carboxaldehydes
24−27.21 On the other hand, 3-bromobenzo[b]thiophene (43)
was converted to 3-methoxybenzo[b]thiophene (44),22 which
was formylated with DMF through 2-lithiation to obtain 2-
carbaldehyde 28.
Antiproliferative Effects and SAR Analysis. All new
analogues were evaluated for antiproliferative activity against 10
human tumor cell lines: KB (originally isolated from
epidermoid carcinoma of the nasopharynx), KB-subline KB-
VIN showing MDR phenotype with overexpression of P-gp,
PC-3 (prostate cancer), A549 (lung carcinoma), HepG2
(hepatocellular carcinoma), HCT-8 (colon adenocarcinoma),
and four breast cancer cell lines (MDA-MD-231, SK-BR-3,
MCF-7, and ZR-75-1) (Table 1). The antiproliferative effects of
the compounds were assessed by the sulforhodamine B (SRB)
assay, and IC50 values were calculated from at least three
independent experiments, each performed in duplicate.
Among the benzo[b]thiophen-3′-yl analogues, all four
hydroxyl derivatives 6−9 showed potent antiproliferative effects
against the tested tumor cell lines, including the P-gp-
overexpressing MDR tumor cell line. In contrast, the
corresponding methoxy derivatives 10−12 were less active.
The position of the hydroxyl group on the B-ring had a strong
influence on activity against HCT-8 tumor cells. Compared
with 6 and 8, hydroxyl analogues 7 and 9, which have a hydroxy
group at the para- or ortho-position, respectively, relative to the
sulfur atom, exhibited more than 5-fold reduced inhibitory
effect against HCT-8, while the activity against the other tumor
cell lines was not noticeably affected. The 3′-hydroxybenzo-
[b]thiophen-2′-yl analogue 14 showed similar cytotoxicity to
the nonhydroxylated parent compound 13 against all tested
tumor cell lines except PC-3, against which 14 displayed 4-fold
reduced activity.
In the case of the 1′-naphthyl derivatives, the insertion of a
hydroxyl group on 15 and 16 generally resulted in slightly
better antiproliferative activity as compared with parent
compound 5. This trend was not seen in the benzo[b]-
thiophene analogues. In particular, the antiproliferative effects
of 16 against HepG2, KB, and KB-VIN were 2−6 times greater
than that of the parent compound 5. In addition, unlike the
benzo[b]thiophene analogues, methoxynaphthalene derivative
17 also inhibited tumor cell growth with IC50 values of 1.5−3.6
μM. These values differed little from those obtained with the
related hydroxyl derivative 16. The hydroxynaphthalen-2′-yl
derivative 19 was generally 10-fold less active than the
hydroxynaphthalen-1′-yl derivatives 15 and 16, while no
differences were found between the activities of 5 (naph-
thalen-1′-yl) and 18 (naphthalen-2′-yl). Overall, in this
flavonoid series, the results indicated that the addition of a
hydroxyl functionality to a bicyclic B-ring system on the first
ring directly connected to the chromene skeleton caused
slightly enhanced activity, although addition of a hydroxyl
group on the second ring not directly connected to the
chromene skeleton generally reduced activity. This SAR study
is summarized in Figure 3.
In this paper, we describe the syntheses of a series of
hydroxyl analogues of 4 and 5 and the bioactivities of these
compounds. We also summarize our molecular modeling
results.
2. RESULTS AND DISCUSSION
Chemistry. Hydroxylated analogues 6−20 (Figure 2) were
prepared from 2-acetyl-4,4,6-triethyl-3-hydroxy-5-methoxycy-
clohexa-2,5-dien-1-one (21) using our prior synthetic method-
ology to produce TEDB (2),12−14 as shown in Scheme 1. A
Claisen−Schmidt condensation of 21 with various
methoxybenzo[b]thiophenecarboxaldehydes 24−28 or com-
mercially available methoxynaphthaldehydes 29−31 provided
the related chalcones 22 in good yield. The resulting chalcones
were cyclized in the presence of catalytic iodine and concd
H2SO4. Demethylation of 23 with BBr3 produced the desired
hydroxyl analogues 6−9, 14−16, and 19. Methoxy analogues
10−12 were obtained as minor products during the
demethylation step. Depending on the reaction conditions,
the methoxy group on the A-ring was removed faster than that
on the B-ring. Compounds 13, 17, 18, and 20 were previously
synthesized.15 The required benzo[b]thiophene aldehydes 24−
27 were constructed as shown in Scheme 2. The condensation
of p-, m-, and o-methoxybenzenethiols (32−34) with α-chloro
ketone, followed by cyclization with polyphosphoric acid at 100
°C, afforded 5-methoxy-3-methylbenzo[b]thiophenes (35)
Scheme 2. Preparation of
Methoxybenzothiophenecarboxaldehydes
a
a
Importantly, all analogues displayed potent antiproliferative
effects against KB-VIN, a MDR KB subline overexpressing P-
gp. The antiproliferative activities of 7−10, 12−14, 16, and 19
against KB-VIN were even greater than those against the parent
Reagents and conditions: (a) CH3COCH2Cl, K2CO3, DMF, rt; (b)
PPA, 100 °C; (c) NBS, (PhCOO)2, CCl4, reflux; (d) HMTA, CHCl3,
reflux, then 50% HOAc, reflux; (e) NaOMe, CuO, CuI, KI, DMF,
reflux; (f) n-BuLi, THF, −78 °C, then DMF.
2380
DOI: 10.1021/jm501859j
J. Med. Chem. 2015, 58, 2378−2389

Journal of Medicinal Chemistry
Article
a
Table 1. Antiproliferative and Antitubulin Activities of Synthesized Analogues against Various Cell Lines
b
IC50 (μM)
tubulin assay
f
g
B-ring
compd
KB
KB-VIN PC-3 A549 Hep-G2 HCT-8 MDA-MB-231 SK-BR3 MCF-7 ZR-75-1 ITA (μM) ICB (%)
d
1
>135
0.08
2.01
2.46
2.98
1.28
6.76
0.07
0.15
1.76
1.51
1.12
11.8
NA
>40
0.16
1.73
3.12
2.12
1.47
NA
>40
0.06
1.78
2.46
1.66
2.04
NA
NT
NT
>40
0.70
0.71
2.21
1.55
NT
NT
NT
8.17
NT
NT
NT
NT
NT
NT
NT
NT
NT
>40
0.50
0.65
4.96
4.80
NT
>40
0.72
2.75
6.44
1.81
NT
NT
NT
7.22
15.2
NT
1.55
NT
NT
3.59
1.42
NT
NT
>40
0.85
11.6
9.31
5.23
NT
NT
NT
9.81
NT
NT
NT
NT
NT
NT
NT
NT
NT
NA
2.0
−
78
c
benzo[b]thiophene
4
0.13
1.29
3.02
1.71
3.01
NA
0.13
1.71
12.2
2.37
10.6
NA
6
2.8
70
7
7.5
32
8
5.5
39
9
4.1
37
e
10
11
12
NA
18.7
NA
NT
NT
NT
NT
NT
2.3
NT
NT
NT
NT
NT
43
NA
NA
NA
NA
16.2
NA
12.1
11.2
19.6
1.55
0.99
0.74
1.67
1.03
19.9
10.8
10.7
9.90
6.61
0.77
0.99
0.45
1.56
1.16
13.7
9.57
8.51
3.81
14.6
1.03
3.04
1.11
2.39
2.58
12.9
4.78
15.8
14.0
12.5
1.55
1.24
0.94
1.48
1.16
14.9
16.8
NA
7.97
NT
c
13
15.7
15.8
2.06
1.63
0.37
1.48
1.55
13.4
13.4
19.0
16.5
1.03
1.34
1.02
3.35
1.29
14.1
12.0
14
12.1
NT
c
naphthalene
5
15
16
0.89
NT
2.4
55
3.4
54
c
17
2.27
NT
2.7
31
c
h
18
NO
NT
NT
1.1
29
19
13.8
8.37
NT
NT
99
c
20
CA-4 (known potent tubulin polymerization inhibitor)
a
KB (epidermoid carcinoma of the nasopharynx), KB-VIN (P-gp-overexpressing MDR subline of KB), PC-3 (prostate cancer), A549 (lung
carcinoma), HepG2 (hepatocellular carcinoma), HCT-8 (human colon adenocarcinoma), MDA-MB-231 (triple-negative breast cancer), SK-BR-3
(estrogen-receptor-negative, progesterone-receptor-negative, HER2-overexpressing breast cancer), MCF-7 (estrogen-receptor-positive, HER2-
b
negative breast cancer), ZR-75-1 (estrogen-receptor-positive, HER2-positive breast cancer). Antiproliferative activity as IC50 values for each cell
c
d
line, the concentration of compound that caused 50% reduction relative to untreated cells determined by the SRB assay. See ref 15. NT, not tested.
e
f
g
NA, not active (IC50 > 10 μg/mL). Inhibition of purified tubulin assembly, EC50 (μM) values of 50% inhibition (ITA). Percent inhibition of 5 μM
h
[3H]colchicine binding to 1 μM tubulin in the presence of 5 μM test compound (ICB). NO, Not obtainable. Partial inhibition was observed and
was maximal with 4 μM compound. Higher concentrations of the compound resulted in the same amount of inhibition observed with 4 μM,
suggesting poor solubility of the compound.
Figure 3. Structure−activity relationship of hydroxylated TEDB-TB analogues. SAR study shown in Table 1 summarized: (1) 4′-position ≈ 6′-
position > 5′-position ≈ 7′-position against HCT-8 and (2) H > OH against PC-3.
KB non-MDR tumor cell line. This fact indicates that these
analogues are not P-gp substrates and could be effective against
tumors expressing the MDR phenotype.
tubulin assembly with an EC50-ITA value of 2.8 μM and
inhibited colchicine binding by 70%. The data for 4 and 6 were
very similar, despite the more than 10-fold difference in their
antiproliferative effects against tumor cells. These results
suggest that both compounds bind to tubulin and inhibit its
polymerization, while 6 may be detoxified in the tumor cells
more rapidly than 4. The disparate results between
antiproliferative and tubulin inhibitory effects were also
observed for other analogues. 4′-Hydroxyl analogue 6 and 6′-
hydroxyl analogue 8 comparably inhibited tumor cell growth,
but the inhibitory effects on tubulin assembly and colchicine
binding of 6 were 2-fold greater than those of 8. These results
prompted us to perform additional investigations on hydroxyl
analogs 6 and 8, as well as 4, using human tumor cell lines. It
Inhibition of Tubulin Polymerization and Colchicine
Binding to Tubulin. Cytotoxic analogues 4−9 and 15−18
were tested for inhibitory effects on tubulin assembly as well as
inhibition of [3H]colchicine binding to tubulin, using highly
purified bovine brain tubulin. Table 1 lists the 50% effective
concentration for inhibiting tubulin assembly (EC50-ITA) and
the percent inhibition of colchicine binding to tubulin (ICB) in
the presence of compounds 4−9 and 15−18, as well as results
obtained with the CS antimicrotubule agent CA-4 as a positive
control. Standard deviations can be found in the Supporting
Information. Analogue 6 with a 4′-hydroxyl strongly inhibited
2381
DOI: 10.1021/jm501859j
J. Med. Chem. 2015, 58, 2378−2389

Journal of Medicinal Chemistry
Article
Figure 4. Effect of compounds on the cell cycle distribution in MDR cells. Chemosensitive KB cell line (A) and its MDR subline (KB-VIN) (B)
were treated with vehicle control (DMSO), PXL (0.02 μM for KB and 3 μM for KB-VIN), 4, or 8 for 6 or 24 h, as indicated. Cell cycle progression
was analyzed by measuring DNA content using a flow cytometer. Cell cycle phases are indicated as G1 (2N), S, G2/M (4N), and sub-G1. PXL
disrupted cell cycle progression in KB cells at 0.02 μM, while no significant effect was seen in KB-VIN cells, even at 3 μM. Accumulation of G2/M
phase as well as S phase cells was observed in both chemosensitive and MDR cells treated with 4 and 8 at the same concentrations in a time-
dependent manner.
should be noted that original natural product 1 showed no
effects on tubulin assembly in a cell-free system (Table 1) or in
KB-VIN cells (data not shown).
antiproliferative assays (Table 1) and indicate that MDR cells
do not efflux 4 and its analogues.
Morphological Variations of Microtubule Dynamics
by Antimicrotubule Agents in Tumor Cells. In the current
study, immunofluorescence visualization of tubulin in PC-3
cells was used to clearly show distinguishable disruptions of
microtubule dynamics upon treatment with different anti-
microtubule agents binding to diverse sites, such as CA-4
binding to the CS, VIN binding to the vinca site, and PXL
binding to the taxane site (Figure 5). In particular, microtubule
morphology in mitotic PC-3 cells treated with 4 was not
identical to those in mitotic cells treated with other tested
antimicrotubule agents, while it was similar to that in interphase
cells treated with CA-4 (Supporting Information, Figure S1).
These results suggest that the mode of action of 4 is to inhibit
tubulin polymerization in the interphase by a mechanism
similar to that of CA-4, but different from CA-4 at the time of
mitotic spindle formation.
Induction of Cell Cycle Arrest at the G2/M Phase in
MDR Tumor Cells. Small molecules showing antimicrotubule
activity, such as colchicine, VIN, and PXL, inhibit cell cycle
progression at the G2/M phase in chemosensitive tumors.
However, these compounds are ineffective against MDR
tumors, because overexpressed ABC transporter(s) efficiently
efflux these compounds.5 To see if the newly synthesized
compounds affected the cell cycle in MDR tumor cells, the cell
cycle progression of treated KB-VIN and parental KB cells was
analyzed by flow cytometry (Figure 4). PXL at 0.02 μM
significantly impacted the cell cycle in KB cells, while PXL at 3
μM did not show any effect in the KB-VIN cells, demonstrating
that PXL is efficiently effluxed from the MDR tumor cell by
overexpressed P-gp. In contrast, accumulation of cells in the
G2/M phase, as well as the S-phase, was observed in both
chemosensitive (KB) and MDR cells (KB-VIN) treated with 4
and its analogues at identical concentrations. Cell cycle
progression was also disrupted by compounds in a time-
dependent manner. These results support the data found in the
Modification Sites in TEDB Skeleton Have Differing
Effects on Microtubule Polymerization. To further under-
stand the effects of the new hydroxyl-modified 4 analogues on
microtubule dynamics and on the cell cycle, dose-dependent
2382
DOI: 10.1021/jm501859j
J. Med. Chem. 2015, 58, 2378−2389

Journal of Medicinal Chemistry
Article
Figure 5. Disruption of microtubule morphology in PC-3 cells by small molecules. Prostate cancer cells (PC-3) were treated for 24 h with
antimicrotubule agents as indicated (DMSO as vehicle control, 0.2 μM compound 4, 0.2 μM CA-4, 5 μM VIN, 0.2 μM PXL). Compounds were
used at equitoxic compound concentrations and examined for disruption of microtubule dynamics within 24 h. Microtubules were visualized by
immunofluorescence staining using monoclonal antibody to α-tubulin (upper panels), and DAPI was used for DNA staining (lower panels). In
control cells treated with DMSO, bipolar spindle formation in mitotic cells (arrow) as well as microtubule (MT) networks in interphase cells
(double arrow) were observed. Treatment with 4 induced depolymerization of cytosolic microtubules as well monopolar (arrow heads) or multipolar
spindle formations in the G2/M-phase. Treatment with CA-4 led to both depolymerization of interphase microtubules and disruption of mitotic
spindles (arrows). PXL inhibited the onset of mitosis and arrested cells at prometaphase with condensed chromosomes (right panels). Characteristic
bundled microtubules were observed in interphase cells treated with PXL, while tubulin paracrystal formation was observed in cells treated with VIN.
Bar, 25 μm. Additional images including colchicine treatment are shown in the Supporting Information (Figure S1).
morphological microtubule defects were analyzed in PC-3 cells
(Figure 6; Supporting Information, Figure S2). Flow cytometric
analysis clearly demonstrated that all hydroxylated 4 analogues
induced significant cell cycle accumulation in the G2/M phase
(Figure 6A). To determine whether the cell cycle arrest was
due to an antimicrotubule effect, cells treated with compound
were analyzed by triple staining with antibodies to α-tubulin
and the mitotic marker Ser10-phosphorylated histone H3 (p-
H3) and with 4′,6-diamidino-2-phenylindole (DAPI) for DNA.
In cells treated with compounds, multiple spindles were seen in
the p-H3 positive cells (Figure 6B,C). The results demon-
strated that the new compounds inhibited normal bipolar
spindle formation and induced cell cycle arrest at the G2/M
phase, probably at prometaphase. However, the impact of 4
analogues on spindle formation was different from that of other
antimicrotubule agents. While multipolar spindles formed
partially in cells treated with test compounds, no spindles
were detected after treatment with CA-4 or PXL (Figure 6B,C
and Supporting Information, Figure S2). While inhibition of
spindle formation corresponded directly to depolymerization of
interphase microtubules in cells treated with CA-4 or with 4, 6,
or 9, disrupted spindle formation was observed in conjunction
with normally polymerized interphase microtubules in cells
treated with 7 or 8. Compound 7 was less active than 8.
Furthermore, analysis of microtubule morphology in the
presence of different doses of 8 demonstrated that the normal
morphology of interphase microtubules in the presence of 8 did
not result from partial inhibition of tubulin polymerization.
Interestingly, stabilized interphase microtubules were observed
in cells treated with a high concentration (20 μM) of 7 or 8.
Thus, compounds 7 and 8 selectively disrupted spindle
formation without depolymerizing interphase microtubules
(Figure 6B,C and Supporting Information, Figure S2). In
contrast, using CA-4, it was not possible to determine a
concentration that was less effective against interphase
microtubule polymerization than spindle formation. At the
concentration of CA-4 that inhibited interphase microtubule
polymerization, accumulated p-H3 (Ser10)-positive cells
showed no spindle formation. These cell-based observations
suggested that compounds 7 and 8, especially 8, are more
effective against the more dynamic spindles than the less
dynamic interphase microtubules and, thus, interact with
tubulin in a manner distinct from that of CA-4 and of 4, 6,
and 9.
We next used molecular docking studies to investigate the
binding of 8 to the tubulin dimer. Our goal was to determine
whether the compound(s) might have distinct binding
properties that could explain the different antimicrotubule
effects of compounds 4 and 8, as well as colchicine, on tumor
cells.
Computer Modeling. The theoretical binding mode of
TEDB-TBs at the CS in the tubulin dimer was investigated by
using entry 1SA0 in the Protein Data Bank as a structure of the
αβ-tubulin dimer. In this crystal structure, N-deacetyl-N-(2-
mercaptoacetyl)colchicine (DAMA-colchicine), a close ana-
logue of colchicine, was used to illustrate its binding to the
CS.23 Colchicine is known to bind at the interface between the
α- and β-subunits, extending toward the β-subunit (Supporting
Information, Figure S3A).24 Loop 7 (T7), strands 8 and 9, as
well as helixes 7 and 8 (H7 and H8), where the β-subunit
connects to loop 5 (T5) of the α-subunit, are critical for
interactions of ligand and protein in the CS. We previously
reported15 that the benzo[b]thiophenyl B-ring of 4 and the
colchicine A-ring showed common pharmacophores acting as
hydrophobic centers. The overlapped 3D model of 4 and
colchicine (Figure 7A) also supported this prior finding,
showing that both rings were oriented in the same direction.
This partial similarity in the 3D model complemented the
similar effects on inhibition of microtubule polymerization in a
cell-free system (Table 1), as well as in cultured cells (Figure
5).
In the tubulin binding assay described above (Table 1), CA-4
inhibited 99% of colchicine binding to tubulin, while 4 or 8
inhibited 78% or 39%, respectively. Thus, we expected that 4
and its hydroxyl analogues, especially 8, would show a different
mode of binding to the CS. An overview of the binding modes
of 4 and 8 in the crystal structure of the αβ-tubulin dimer did
show overlap with DAMA-colchicine (Figure 7B). However, as
expected, a detailed docking model of 4 (Figure 8A) revealed a
considerable difference from that of DAMA-colchicine
2383
DOI: 10.1021/jm501859j
J. Med. Chem. 2015, 58, 2378−2389

Journal of Medicinal Chemistry
Article
Figure 6. Distinct effects on microtubule morphology by structural modification of compounds. (A) Effect of compounds on cell cycle distribution in
PC-3 cells. PC-3 cells were treated for 24 h with 2 μM 4, 6, 7, 8, or 9 or vehicle control (DMSO), as indicated. Cell cycle progression was analyzed
by flow cytometric analysis. All analogues efficiently induced accumulation of G2/M phase cells. (B) Dose-dependent effects of compounds on
microtubule morphology in PC-3 cells. PC-3 cells were treated for 24 h with 4, 6, 7, 8, or 9 at 0.2, 2, or 20 μM. DMSO was used with the control
cells. CA-4 was used at 0.002 or 20 μM. Cells were stained with antibodies to α-tubulin (green) and Ser10-phosphorylated histone H3 (p-H3, red) as
a mitotic marker and with DAPI for DNA (blue). Stacked and merged confocal images are presented. TEDB-TB analogues induced cell cycle arrest
at prometaphase (p-H3-positive) by disrupting bipolar spindle formation. Immature multipolar spindles (white arrow heads) were formed in the
cells treated with analogues but were undetectable after CA-4 treatment. Dose-dependent disruption of both interphase microtubules and spindles
was observed in cells treated with 4, 6, 9, or CA-4. In contrast, in cells treated with 7 or 8, normal interphase microtubules and disrupted spindles
were visualized (yellow arrows). Bar, 25 μm. (C) Higher magnification views. Bar, 10 μm. Additional images are shown in the Supporting
Information (Figure S2).
(Supporting Information, Figure S3A), and 8 also showed a
distinct binding mode (Figure 8B−D and Supporting
Information, Figure S3B). The benzo[b]thiophenyl B-ring of
4 was located in the pocket on β-tubulin. Accordingly, any
substituents on this ring might directly affect the binding
affinity with amino acid(s) on β-tubulin and the shape of the
binding pocket. The unsaturated A-ring of 4 was positioned
toward α-tubulin, with the C-6 ethyl group forced into the open
space between α- and β-tubulin, while the C-8 geminal ethyl
groups were held entirely in β-tubulin. While the 3-methoxy
group on colchicine’s A-ring generates an important interaction
with the side chain of Cys241 in β-tubulin (βCys241) (note
that numbering of β-tubulin residues follows the modeling
convention used in ref 23 rather than the actual sequence
numbers) (Supporting Information, Figure S3A),23 neither of
the TEDB-TB analogues interacted with βCys241. Instead, the
docked model of analogue 4 showed H-bonds with the side
chains of αAsn101, αSer178, and αThr179 (Figure 8A). With
analogue 8, the side chain of βVal238 formed an H-bond with
the C-6′ hydroxyl group; however, the H-bond found between
the C-7 oxygen of 4 and αAsn101 was not present (Figure 8B−
D and Supporting Information, Figure S3B). We made the
following conclusions based on the docking models of 4 and 8:
(1) Bioactive TEDB-TBs 4 and 8 have two oxygen atoms on C-
4 and C-5 in the chromone skeleton, and the C-4 and C-5
oxygen atoms form H-bonds with αSer178 and αThr179,
respectively. (2) The H-bond between αAsn101 and the C-7
oxygen atom of 4 could be a critical factor for inhibition of
tubulin polymerization. (3) The force of the H-bond between
βVal238 and the C-6′ hydroxyl group of 8 could prevent
formation of the H-bond with αAsn101 observed in 4.
2384
DOI: 10.1021/jm501859j
J. Med. Chem. 2015, 58, 2378−2389

Journal of Medicinal Chemistry
Article
Figure 7. Predicted docking models for 4 and 8 binding to tubulin. (A) Structures and 3D models of 4 (blue skeleton) and DAMA-colchicine (gray
skeleton with oxygen in red, nitrogen in blue, and sulfur in yellow). (B) Docking model of 4 (blue skeleton), 8 (brown skeleton), and DAMA-
colchicine (sphere in 3D with carbon in gray, proton in white, oxygen in red, nitrogen in blue, and sulfur in yellow) in the CS (yellow circle) of the
tubulin crystal structure (PDB ID: 1SA0), shown as a ribbon diagram.
Figure 8. Predicted docking models for 4 and 8 binding in the CS. Crystal structures (PDB ID: 1SA0) of α- (white) and β-tubulin (red) are shown
as ribbon diagrams. H-bonds calculated to be less than 3 Å between protein and compounds are represented by dashed lines. Docking models of
compounds (gray skeleton with oxygen in red and sulfur in yellow) 4 (A) and 8 (B) in the CS are shown. Superimposition of docked compound 4
shows H-bonds with the side chains of αAsn101 (dashed blue circle) and αSer178 and αThr179 (dashed green circle). In contrast, with 8, there is a
H-bond between 8 and the side chain of βVal238 (dashed white circle), and the H-bond between 8 and the side chain of αAsn101 is absent. (C)
Docking mode of 8 in the CS. Superimposition of docked compound 8 shows H-bonds between 8 and αSer178 and αThr179 in α-tubulin (αSer178,
αThr179) and between 8 and βVal238. (D) Comparison of the docking mode of 8 (brown) with that of 4 (blue) in CS. Superimposition of docked
compounds 4 and 8 shows conserved H-bonds with αSer178 and αThr179 (dashed green circle). An additional H-bond between the C-7 oxygen of
4 and αAsn101 (dashed blue circle) was also observed. In contrast, a H-bond between the C-6′ hydroxyl group of 8 and the side chain of βVal238
(white circle) was unambiguous, while no H-bond occurred with αAsn101.
The models for the binding of TEDB-TBs to tubulin show
great dependence on amino acid residues of the α-tubulin,
rather than the β-tubulin, subunit, which differs from other CS
agents and, thus, could explain some of the properties of the
TEDB-TBs. All four hydroxyl benzo[b]thiophen-3-yl analogues
6−9 and lead compound 4 bound in the same place on tubulin
with similar docking structures, although the 7′-OH analogue 9
was slightly shifted from the docking pattern of the other
compounds (Figure 9A,B). This difference indicated that the
binding pocket around the 7′-position either did not have
enough space to accommodate a hydroxyl group or generated
an electronic disruption.
Unexpectedly, the distance between the A-ring carbonyl
oxygen on 9 and the nitrogen on αAsn101 (N−O distance)
was not large (3.076 Å), despite the slight shift in the docking
mode of 9 as described above. Furthermore, analogue 6
displayed the smallest N−O distance (2.563 Å), although an
unsuitable N−H···O angle (111.6°) appeared to prevent H-
bond formation. For 4, the N−O distance was 3.091 Å, which is
more than 0.8 Å shorter than that of 8 (Figure 9D).
2385
DOI: 10.1021/jm501859j
J. Med. Chem. 2015, 58, 2378−2389

Journal of Medicinal Chemistry
Article
Figure 9. Hydrogen bonds between 4 analogues and tubulin in the CS. Structures (A) and conformations (B) of 4 and analogues are presented. The
same docking pattern is shown among all compounds, except 9. (C) H-bond between C-7 oxygen of compounds and αAsn101. A H-bond calculated
as less than 3 Å occurred in 4 and 6. (D) Comparison of distance between oxygen on compound and nitrogen on αAsn101. The N−O distances for
4 (3.091 Å) and 8 (3.963 Å) differed by over 0.8 Å.
Accordingly, analogues 4 and 6 were located closer to αAsn101
than analogues 7 and 8 (Figure 9C). This result might explain
why analogues 4 and 6 exhibited greater tubulin depolymeriza-
tion activity than analogues 7 and 8. The distinction between
the H-bonding modes of 4 and 8 might be relevant to their
distinct antimicrotubule activities in cultured cells. Accordingly,
like CA-4, compounds 4 and 6 disrupted interphase micro-
tubule dynamics as well as spindle formation, while interphase
microtubules were intact in cells treated with 7 and 8 at the
effective dose that induced cell cycle arrest by inhibiting spindle
formation (Figure 6).
Finally, it should be noted that the C-7′ methoxy group in
analogue 12 appeared to be located similarly to, but about 2 Å
inside, an A-ring methoxy group on colchicine. Thus, the 7′-
OMe group might be too close to the βVal238 and βThr239
positions and could disturb the binding affinity (Supporting
Information, Figure S4).
In summary, hydroxyl- or methoxy-substituted TEDB-TB
analogues 6−20 were synthesized and evaluated for in vitro
antiproliferative activity, for effects on cell cycle progression,
and for interactions with purified tubulin. Hydroxyl naph-
thalene derivatives 15 and 16 inhibited growth of multiple
tumor cell lines more strongly than their parent compound 5.
All hydroxyl analogues showed potent antiproliferative activity,
while methoxy analogues were less active. Active analogues
induced cell cycle arrest at the G2/M phase, especially at the
prometaphase by disrupting spindle formation. All cytotoxic
hydroxyl analogues targeted tubulin. Molecular modeling
indicated that the hydroxyl benzo[b]thiophene analogues
interacted with the CS in distinct binding modes. Disruption
of interphase microtubules by the benzo[b]thiophene ana-
logues was altered by modification of the parent compound
with hydroxyl groups in a site-dependent manner.
attributable to the modification of the benzo[b]thiophene
moiety with hydroxyl groups. In particular, αAsn101, αSer178,
and αThr179 and βVal238 contributed to stable H-bonds with
TEDB-TB analogues. Compound 4, which disrupted both
interphase microtubules and mitotic spindle formation, and
compound 8, which disrupted only spindle formation, formed
H-bonds with only α-tubulin or with both α- and β-tubulin,
respectively. Thus, we speculate that interphase microtubule
dynamics are less affected or even not disrupted at all when a
TEDB-TB analogue/ligand, such 8, formed several H-bonds
with both α- and β-tubulin.
While clinically successful tubulin inhibitors target both the
mitotic spindle and interphase microtubules in chemosensitive
cancer cells, some of our new analogues effectively inhibited
bipolar spindle formation without interfering with interphase
microtubule dynamics in both chemosensitive and MDR cancer
cells. Therefore, these TEDB analogues might be promising
candidates leading to a new chemotherapeutic drug targeted at
tubulin and also could possibly overcome MDR obstacles, as
well as reducing undesirable effects in slowly growing normal
cells.
3. EXPERIMENTAL SECTION
All chemicals and solvents were used as purchased. All melting points
were measured on a Fisher-Johns melting point apparatus and are
1
reported without correction. H and 13C NMR spectra were recorded
on a Varian Gemini 2000 (300 MHz) or Varian Inova (400 MHz)
NMR spectrometer with TMS as the internal standard. All chemical
shifts are reported in ppm. NMR spectra are referenced to the residual
solvent peak, chemical shifts δ are in ppm, and apparent scalar
coupling constants J are in hertz. Mass spectroscopic data were
obtained on a Shimadzu LCMS-IT-TOF instrument. Analytical thin-
layer chromatography (TLC) was carried out on Merck precoated
aluminum silica gel sheets (Kieselgel 60 F-254). Biotage Flash or Isco
Companion systems were used for flash chromatography. All target
compounds were characterized and determined to be at least >95%
Computational docking models suggested that all TEDB-TB
analogues bound to the tubulin heterodimer at the CS, but
considerable differences were observed in H-bond formations
with CS amino acids. This binding mode diversity could be
1
pure by H NMR and analytical HPLC.
General Synthetic Procedures for 22. A solution of 21 in
EtOH−50% aq KOH (1:1, v/v) and an appropriate aromatic aldehyde
2386
DOI: 10.1021/jm501859j
J. Med. Chem. 2015, 58, 2378−2389

Journal of Medicinal Chemistry
Article
(excess) was stirred at room temperature. After the reaction was
complete, as judged by TLC analysis, the mixture was poured into ice-
cold 1 N HCl and then extracted with CH2Cl2. The extract was
washed with brine, dried over Na2SO4, and concentrated in vacuo. The
residue was chromatographed on silica gel with CH2Cl2−hexane as
eluent to afford the target compound 22 in 78%−95% yield (based on
recovery of starting material).
General Synthetic Procedures for 4−20. Compound 22 was
dissolved in DMSO containing 1% H2SO4, then I2 (0.1 mol equiv) was
added. The mixture was heated at 90 °C for 1 h. The reaction mixture
was quenched with ice-cold aqueous 10% Na2S2O3 and extracted three
times with EtOAc. The combined organic layers were washed with
brine, dried over Na2SO4, and concentrated in vacuo. The residue was
purified by silica gel chromatography with EtOAc−hexane as eluent to
afford crude compound 23, which was dissolved in anhydrous CH2Cl2.
BBr3 (3 mol equiv, 1.0 M solution in CH2Cl2) was added to the
solution at 0 °C, which was allowed to warm to rt and stirred
overnight. After addition of water, the reaction mixture was extracted
three times with CH2Cl2. The combined organic layers were washed
with brine, dried over Na2SO4, and concentrated in vacuo. The
residues were chromatographed on silica gel, eluting with EtOAc−
hexane (1:4) to obtain analogues 6−9, 14−16, and 19, as well as 10−
12, 17, and 20 as minor products.
2-(4′-Hydroxybenzo[b]thiophen-3′-yl)-6,8,8-triethyldesmos-
dumotin B (6). 1H NMR (300 MHz, CDCl3, δ): 13.12 (s, 1H,
chelated-OH), 7.75 (s, 1H, 2′-H), 7.61 (d, 1H, J = 8.0 Hz, 5′ or 7′-H),
7.31 (dd, 1H, J = 7.8 and 8.0 Hz, 6′-H), 6.99 (s, 1H, 3-H), 6.90 (d, 1H,
J = 7.8 Hz, 5′ or 7′-H), 2.47 (q, 2H, J = 7.3 Hz, 6-CH2CH3), 2.21−
2.10 (m, 2H, 8-CH2CH3), 2.03−1.91 (m, 2H, 8-CH2CH3), 1.06 (t,
3H, J = 7.3 Hz, 6-CH2CH3), 0.68 (t, 6H, J = 7.3 Hz, 8-CH2CH3 × 2).
HRMS (m/z): [M + H]+ calcd for C23H23O5S, 411.1261; found,
411.1249.
2-(5′-Hydroxybenzo[b]thiophen-3′-yl)-6,8,8-triethyldesmos-
dumotin B (7). 1H NMR (300 MHz, CDCl3, δ): 13.10 (s, 1H,
chelated-OH), 8.08 (s, 1H, 2′-H), 7.80 (d, 1H, J = 8.7 Hz, 7′-H), 7.57
(d, 1H, J = 2.3 Hz, 4′-H), 7.10 (dd, 1H, J = 8.7 and 2.3 Hz, 6′-H), 6.95
(s, 1H, 2-H), 5.55 (s, 1H, 5′-OH), 2.47 (q, 2H, J = 7.3 Hz, 6-
CH2CH3), 2.34−2.23 (m, 2H, 8-CH2CH3), 2.06−1.96 (m, 2H, 8-
CH2CH3), 1.06 (t, 3H, J = 7.3 Hz, 6-CH2CH3), 0.71 (t, 6H, J = 7.3
Hz, 8-CH2CH3 × 2). HRMS (m/z): [M + H]+ calcd for C23H23O5S,
411.1261; found, 411.1258.
chelated-OH), 8.05 (s, 1H, 2′-H), 7.82 (d, 1H, J = 8.9 Hz, 7′-H), 7.49
́
(t, 1H, J = 2.3 Hz, 4-H), 7.15 (dd, 1H, J = 8.9 and 2.3 Hz, 6′-H), 6.89
(s, 1H, 3-H), 3.91 (s, 3H, 5′-OCH3), 2.47 (q, 2H, J = 7.5 Hz, 6-
CH2CH3), 2.34−2.22 (m, 2H, 8-CH2CH3), 2.07−1.96 (m, 2H, 8-
CH2CH3), 1.06 (t, 3H, J = 7.5 Hz, 6-CH2CH3), 0.71 (t, 6H, J = 7.5
Hz, 8-CH2CH3 × 2). HRMS (m/z): [M + H]+ calcd for C24H25O5S,
425.1417; found, 425.1402.
2-(7′-Methoxybenzo[b]thiophen-3′-yl)-6,8,8-triethyldes-
mosdumotin B (12). 1H NMR (300 MHz, CDCl3, δ): 13.10 (s, 1H,
chelated-OH), 8.05 (s, 1H, 2′-H), 7.68 (d, 1H, J = 8.2 Hz, 7′-H), 7.51
(dd, 1H, J = 7.8 and 8.2 Hz, 6′-H), 6.93 (s, 1H, 3-H), 6.93 (d, 1H, J =
7.8 Hz, 5′-H), 4.06 (s, 3H, 7′-OCH3), 2.47 (q, 2H, J = 7.3 Hz, 6-
CH2CH3), 2.32−2.21 (m, 2H, 8-CH2CH3), 2.06−1.94 (m, 2H, 8-
CH2CH3), 1.06 (t, 3H, J = 7.3 Hz, 6-CH2CH3), 0.71 (t, 6H, J = 7.3
Hz, 8-CH2CH3 × 2). HRMS (m/z): [M + H]+ calcd for C24H25O5S,
425.1417; found, 425.1420.
2-(3′-Hydroxybenzo[b]thiophen-2′-yl)-6,8,8-triethyldesmos-
1
dumotin B (14). H NMR (300 MHz, CDCl3, δ): 12.81 (s, 1H,
chelated-OH), 11.65 (br s, 1H, 3′-OH), 8.23−8.17 (m, 1H), 8.04 (s,
1H, 3-H), 7.84−7.79 (m, 1H), 7.59−7.53 (m, 2H), 2.51 (q, 2H, J = 7.5
Hz, 6-CH2CH3), 2.35−2.23 (m, 2H, 8-CH2CH3), 2.12−2.00 (m, 2H,
8-CH2CH3), 1.09 (t, 3H, J = 7.5 Hz, 6-CH2CH3), 0.71 (t, 6H, J = 7.5
Hz, 8-CH2CH3 × 2). HRMS (m/z): [M + H]+ calcd for C23H23O5S,
411.1261; found, 411.1251.
2-(4′-Hydroxynaphthalen-1′-yl)-6,8,8-triethyldesmosdumo-
1
tin B (16). H NMR (300 MHz, CDCl3, δ): 13.25 (s, 1H, chelated-
OH), 8.41−8.36 (m, 1H, naphthalenyl-H), 7.98−7.92 (m, 1H,
naphthalenyl-H), 7.65−7.58 (m, 2H, naphthalenyl-H), 7.56 (d, 1H, J
= 8.0 Hz, naphthalenyl-H), 6.97 (d, 1H, J = 8.0 Hz, naphthalenyl-H),
6.79 (s, 1H, 3-H), 6.70 (br s, 1H, 4′-OH), 2.50 (q, 2H, J = 7.4 Hz, 6-
CH2CH3), 2.26−2.16 (m, 2H, 8-CH2CH3), 1.99−1.88 (m, 2H, 8-
CH2CH3), 1.07 (t, 3H, J = 7.4 Hz, 6-CH2CH3), 0.73 (t, 6H, J = 7.4
Hz, 8-CH2CH3 × 2). HRMS (m/z): [M + H]+ calcd for C25H25O5,
405.1696; found, 405.1680.
2-(6′-Hydroxynaphthalen-2′-yl)-6,8,8-triethyldesmosdumo-
tin B (19). 1H NMR (300 MHz, 3%CD3OD/CDCl3, δ): 13.21 (s, 1H,
chelated-OH), 8.23 (br s, 1H, naphthalenyl-H), 7.87 (d, 1H, J = 8.8
Hz, naphthalenyl-H), 7.80 (d, 1H, J = 8.8 Hz, naphthalenyl-H), 7.72
(dd, 1H, J = 8.8 and 1.9 Hz, naphthalenyl-H), 7.26−7.19 (m, 2H,
naphthalenyl-H), 6.98 (s, 1H, 3-H), 2.48 (q, 2H, J = 7.4 Hz, 6-
CH2CH3), 2.35−2.24 (m, 2H, 8-CH2CH3), 2.13−2.01 (m, 2H, 8-
CH2CH3), 1.06 (t, 3H, J = 7.4 Hz, 6-CH2CH3), 0.70 (t, 6H, J = 7.4
Hz, 8-CH2CH3 × 2). HRMS (m/z): [M + H]+ calcd for C25H25O5,
405.1696; found, 405.1679.
Antiproliferative Activity Assay. All stock cell lines were grown
in T-75 flasks at 37 °C with 5% CO2 in air. Freshly trypsinized cell
suspensions were seeded in 96-well microtiter plates at densities of
4000−12 000 cells per well (based on the doubling time of the cell
line) with compounds. The highest concentration of DMSO in the
cultures (0.1% v/v) was without effect on cell growth under the
culture conditions used. After 72 h in culture with test compounds,
attached cells were fixed with 50% trichloroacetic acid and then stained
with 0.04% sulforhodamine B. After solubilizing the protein-bound dye
with 10 mM Tris base, absorbance at 515 nm was measured using a
microplate reader (ELx800, BioTek) with Gen5 software (BioTek).
The mean IC50 is the concentration of agent that reduced cell growth
by 50% compared with vehicle (DMSO) control under the
experimental conditions used and is the average from at least three
independent experiments with duplicate samples. All values presented
in Table 1 are statistically significant. The following human tumor cell
lines were used in the assay: A549 (lung carcinoma), HepG2
(hepatocellular carcinoma), HCT-8 (colon adenocarcinoma), KB
(originally isolated from epidermoid carcinoma of the nasopharynx),
KB-VIN (VIN-resistant KB subline showing MDR phenotype by
overexpressing P-gp), MCF-7 (estrogen-receptor-positive, HER2-
negative breast cancer), PC-3 (androgen-insensitive prostate cancer),
SK-BR-3 (estrogen-receptor-negative, progesterone-receptor-negative,
HER2-overexpressing breast cancer), ZR-75-1 (estrogen-receptor-
positive, HER2-overexpressing breast cancer). All cell lines were
obtained from the Lineberger Comprehensive Cancer Center (UNC-
2-(6′-Hydroxybenzo[b]thiophen-3′-yl)-6,8,8-triethyldesmos-
dumotin B (8). 1H NMR (300 MHz, CDCl3, δ): 13.11 (s, 1H,
chelated-OH), 7.94 (d, 1H, J = 8.8 Hz, 4′-H), 7.88 (s, 1H, 2′-H), 7.41
(d, 1H, J = 2.3 Hz, 7′-H), 7.12 (dd, 1H, J = 8.8 and 2.3 Hz, 5′-H), 6.90
(s, 1H, 3-H), 5.62 (s, 1H, 6′-OH), 2.48 (q, 2H, J = 7.4 Hz, 6-
CH2CH3), 2.34−2.21 (m, 2H, 8-CH2CH3), 2.05−1.94 (m, 2H, 8-
CH2CH3), 1.06 (t, 3H, J = 7.4 Hz, 6-CH2CH3), 0.71 (t, 6H, J = 7.4
Hz, 8-CH2CH3 × 2). HRMS (m/z): [M + H]+ calcd for C23H23O5S,
411.1261; found, 411.1258.
2-(7′-Hydroxybenzo[b]thiophen-3′-yl)-6,8,8-triethyldesmos-
dumotin B (9). 1H NMR (300 MHz, CDCl3, δ): 13.12 (s, 1H,
́
chelated-OH), 8.07 (s, 1H, 2-H), 7.68 (d, 1H, J = 8.1 Hz, 4′-H), 7.43
(dd, 1H, J = 7.8 and 8.1 Hz, 5′-H), 6.94 (s, 1H, 3-H), 6.92 (d, 1H, J =
́
7.8 Hz, 6-H), 6.10 (s, 1H, 7′-OH), 2.48 (q, 2H, J = 7.4 Hz, 6-
CH2CH3), 2.34−2.12 (m, 2H, 8-CH2CH3), 2.08−1.96 (m, 2H, 8-
CH2CH3), 1.06 (t, 3H, J = 7.4 Hz, 6-CH2CH3), 0.72 (t, 6H, J = 7.4
Hz, 8-CH2CH3 × 2). HRMS (m/z): [M + H]+ calcd for C23H23O5S,
411.1261; found, 411.1250.
2-(4′-Methoxybenzo[b]thiophen-3′-yl)-6,8,8-triethyldes-
mosdumotin B (10). 1H NMR (300 MHz, CDCl3, δ): 13.21 (s, 1H,
chelated-OH), 7.66 (s, 1H, 2′-H), 7.53 (d, 1H, J = 8.0 Hz, 7′-H), 7.42
(t, 1H, J = 8.0 Hz, 6′-H), 6.88 (d, 1H, J = 8.0 Hz, 5′-H), 6.71 (s, 1H, 3-
H), 3.85 (s, 3H, 4′-OCH3), 2.47 (q, 2H, J = 7.4 Hz, 6-CH2CH3),
2.19−2.09 (m, 2H, 8-CH2CH3), 1.94−1.83 (m, 2H, 8-CH2CH3), 1.06
(t, 3H, J = 7.4 Hz, 6-CH2CH3), 0.68 (t, 6H, J = 7.4 Hz, 8-CH2CH3 ×
2). HRMS (m/z): [M + H]+ calcd for C24H25O5S, 425.1417; found,
425.1415.
2-(5′-Methoxybenzo[b]thiophen-3′-yl)-6,8,8-triethyldes-
mosdumotin B (11). 1H NMR (300 MHz, CDCl3, δ): 13.10 (s, 1H,
2387
DOI: 10.1021/jm501859j
J. Med. Chem. 2015, 58, 2378−2389

Journal of Medicinal Chemistry
Article
CH) or from ATCC (Manassas, VA), except KB-VIN, which was a
generous gift of Prof. Y.-C. Cheng (Yale University). Cells were
cultured in RPMI-1640 medium supplemented with 2 mM L-
glutamine and 25 mM HEPES (HyClone), supplemented with 10%
heat-inactivated fetal bovine serum (HyClone), 100 μg/mL
streptomycin, 100 IU/mL penicillin, and 0.25 μg/mL amphotericin
B (Cellgro). MDR stock cells (KB-VIN) were maintained in the
presence of 100 nM VIN.
Tubulin Assays. Inhibitory effects of compounds against electro-
phoretically homogeneous bovine brain tubulin were as described
previously.25,26 Tubulin assembly assay mixtures contained 1.0 mg/mL
(10 μM) tubulin and varying compound concentrations and were
preincubated for 15 min at 30 °C without guanosine 5′-triphosphate
(GTP). The samples were placed on ice, and 0.4 mM GTP was added.
Reaction mixtures were transferred to 0 °C cuvettes, and turbidity
development was followed for 20 min at 30 °C following a rapid
temperature jump. Compound concentrations as EC50 values that
inhibited the increase in turbidity by 50% relative to a control sample
were determined. In colchicine inhibition assays, tubulin (1.0 μM) was
incubated with 5.0 μM [3H]colchicine and 5.0 μM test compound at
37 °C for 10 min, when about 40−60% of maximum colchicine
binding occurs in control samples.
Cell Cycle Analysis. Distribution of cells in the cell cycle was
evaluated by measurement of the cellular DNA content by propidium
iodide (PI) (BD Biosciences) staining. Briefly, cells were seeded in 12-
well culture plates 24 h prior to treatment with compounds. PXL was
used at 0.02 or 3 μM for chemosensitive (KB) or MDR (KB-VIN)
cells, respectively. Both KB and KB-VIN cells were treated with 0.5
μM 4 or 2 μM 8, and vehicle (DMSO) was used as a control. For PC-
3 cells, compounds 4 or 6−9 were used at 2 μM. After a 24 h
treatment, supernatants and trypsinized cells were collected together,
followed by centrifugation for 5 min at 1000 rpm. The pellet was
resuspended with PBS and fixed in 70% EtOH overnight at −20 °C,
followed by staining with PI overnight at 4 °C. Stained cells were
analyzed by flow cytometry (FACS LSRII, BD Biosciences).
Experiments were repeated a minimum of three times.
Immunofluorescence Staining. PC-3 cells were grown on an 8-
well chamber slide (Lab-Tech) for 24 h prior to treatment with
compound at equicytotoxic compound concentrations. The equi-
cytotoxic compound concentrations were determined on the basis of
their IC50 and/or the effective concentration used for cell cycle analysis
as follows: 0.2 μM 4, 0.2 μM CA-4, 5 μM VIN, 0.2 μM PXL, and 1 μM
colchicine (Figure 5 and Supporting Information, Figure S1). PC-3
cells were also treated with different doses of compounds 4 or 6−9 at
0.2, 2, and 20 μM. CA-4 was used at 0.002, 0.02, and 20 μM (Figure 6
and Supporting Information, Figure S2). After treatment of cells with
agent for 24 h, cells were fixed with 4% paraformaldehyde in PBS and
permeabilized with 0.5% Triton X-100 in PBS. Fixed cells were labeled
with mouse monoclonal antibody to α-tubulin (B5-1-2, Sigma) and
rabbit IgG to Ser10-phosphorylated histone H3 (p-H3) (#06570,
EMD Millipore), followed by FITC-conjugated antibody to mouse
IgG (Sigma) and Alexa Fluor 549-conjugated antibody to rabbit IgG
(Life Technologies). Nuclei were labeled with DAPI (Sigma).
Fluorescently labeled cells were observed using an Olympus BX61
fluorescence microscope (Olympus), and images were captured by an
ORCA-R2 digital camera (Hamamatsu Photonics) with Volocity 5.5.2
software (PerkinElmer Inc.). Microtubules, p-H3, and DNA were also
detected using a confocal microscope (Zeiss, LSM700) with ZEN
(black edition) software (Zeiss). Confocal images were stacked and
merged using ZEN (black edition) software. Experiments were
repeated at least twice for each compound at each concentration.
Final images were prepared using Adobe Photoshop CS3.
quantum chemically optimized structures of ligands were used as initial
structures. The structural optimizations of ligands were carried out by
B3LYP/6-311+G(df,p) using Gaussian 09, Revision B.01.29
ASSOCIATED CONTENT
* Supporting Information
■
S
Synthetic procedures of benzo[b]thiophene aldehydes 24−28,
standard deviations of cytotoxicity and antitubulin activity,
effects of compounds on microtubule dynamics in PC-3 cells,
and the docking models of 12 and DAMA−colchicine. These
materials are available free of charge via the Internet at http://
pubs.acs.org.
AUTHOR INFORMATION
Corresponding Authors
*K.N.G.: phone, +81-76-264-6305; e-mail, kngoto@p.
kanazawa-u.ac.jp.
■
*M.G.: phone, +1-919-843-6325; e-mail, goto@med.unc.edu.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We wish to thank the Microscopy Service Laboratory (UNC−
CH) for their expertise in fluorescence microscopy and
confocal microscopy as well as Dr. Leaf Huang (UNC-CH)
for use of his flow cytometer. In addition, we appreciate critical
comments, suggestions, and editing of the manuscript by Dr.
Susan L. Morris-Natschke (UNC-CH). This study was
supported by a Grant-in-Aid from the Ministry of Education,
Culture, Sports, Science and Technology (MEXT KAKENHI,
Japan) awarded to K.N.G. (Grant Number 25293024 &
25670054) and A.O. (Grant Number 23790137), by a grant
from Terumo Life Science Foundation awarded to K.N.G., by
NIH grant CA177584 from the National Cancer Institute
awarded to K.H.L., and by a grant from the Junior Faculty
Development Awards (UNC) as well as the University
Research Council (UNC) awarded to M.G.
ABBREVIATIONS USED
■
CA-4, combretastatin A-4; CSA, colchicine site agent; CS,
colchicine site; DAMA-colchicine, N-deacetyl-N-(2-
mercaptoacetyl)colchicine; DAPI, 4′,6-diamidino-2-phenylin-
dole; DesB, desmosdumotin B; DMF, N,N-dimethylforma-
mide; H-bond, hydrogen bond; IC50, concentration that
inhibits 50% human tumor cell growth; ICB, inhibition of
colchicine binding; ITA, inhibition of tubulin assembly; MDR,
multidrug resistance; PBS, phosphate-buffered saline; P-gp, P-
glycoprotein; PXL, paclitaxel; SAR, structure−activity relation-
ship; TEDB, 6,8,8-triethyldesmosdumotin B; TEDB-TB, 2-
(benzo[b]thiophene-3′-yl)-6,8,8-triethyldesmosdumotin B;
VIN, vincristine
REFERENCES
■
(1) Dumontet, C.; Jordan, M. A. Microtubule-binding agents: A
dynamic field of cancer therapeutics. Nat. Rev. Drug Discovery 2010, 9,
790−803.
Computer Modeling. Three-dimensional (3D) structures of
tubulin-ligand complexes were predicted by GOLD 5.1 software27
with default settings. The 3D structure of human tubulin (TUBA1A
and TUBB2B) used in this study was constructed from a Protein Data
Bank entry (PDB ID: 1SA0).23 Missing hydrogen atoms in the crystal
structure were computationally added by Hermes.28 The center of the
active site was defined as the center of the ligand in 1SA0, and the
active site radius was set to 10.0 Å. For the docking calculations, the
(2) Risinger, A. L.; Giles, F. J.; Mooberry, S. L. Microtubule dynamics
as a target in oncology. Cancer Treat. Rev. 2009, 35, 255−261.
(3) Jordan, M. A.; Wilson, L. Microtubules as a target for anticancer
drugs. Nat. Rev. Cancer 2004, 4, 253−265.
(4) Lu, Y.; Chen, J.; Xiao, M.; Li, W.; Miller, D. D. An overview of
tubulin inhibitors that interact with the colchicine binding site. Pharm.
Res. 2012, 29, 2943−2971.
2388
DOI: 10.1021/jm501859j
J. Med. Chem. 2015, 58, 2378−2389

Journal of Medicinal Chemistry
Article
(5) Eckford, P. D.; Sharom, F. J. ABC efflux pump-based resistance to
chemotherapy drugs. Chem. Rev. 2009, 109, 2989−3011.
(6) Fojo, T.; Menefee, M. Mechanisms of multidrug resistance: The
potential role of microtubule-stabilizing agents. Ann. Oncol. 2007, 18,
3−8.
(7) Yeh, J. J.; Hsu, W. H.; Wang, J. J.; Ho, S. T.; Kao, A. Predicting
chemotherapy response to paclitaxel-based therapy in advanced
nonsmall-cell lung cancer with P-glycoprotein expression. Respiration
2003, 70, 32−35.
(8) Leonard, G. D.; Fojo, T.; Bates, S. E. The role of ABC
transporters in clinical practice. Oncologist 2003, 8, 411−424.
(9) Ling, V. Multidrug resistance: Molecular mechanisms and clinical
relevance. Cancer Chemother. Pharmacol. 1997, 40, S3−8.
(10) Kavellaris, M. Microtubules and resistance to tubulin-binding
agents. Nat. Rev. Cancer 2010, 10, 194−204.
(11) Stengel, C.; Newman, S. P.; Leese, M. P.; Potter, B. V.; Reed, M.
J.; Purohit, A. Class III beta-tubulin expression and in vitro resistance
to microtubule targeting agents. Br. J. Cancer 2010, 102, 316−324.
(12) Nakagawa-Goto, K.; Bastow, K. F.; Wu, J. H.; Tokuda, H.; Lee,
K. H. Total synthesis and bioactivity of unique flavone desmosdumo-
tin B and its analogs. Bioorg. Med. Chem. Lett. 2005, 15, 3016−3019.
(13) Nakagawa-Goto, K.; Bastow, K. F.; Chen, T. H.; Morris-
Natschke, S. L.; Lee, K. H. Antitumor agents 260. New
desmosdumotin B analogues with improved in vitro anticancer
activity. J. Med. Chem. 2008, 51, 3297−3303.
Day, B. W.; Hamel, E. Structure−activity analysis of the interaction of
curacin A, the potent colchicine site antimitotic agent, with tubulin and
effects of analogs on the growth of MCF-7 breast cancer cells. Mol.
Pharmacol. 1998, 53, 62−76.
(27) Jones, G.; Willett, P.; Glen, R. C. Molecular recognition of
receptor sites using a genetic algorithm with a description of
desolvation. J. Mol. Biol. 1995, 245, 43−53.
(28) Hermes 5.1; CCDC Software Ltd.: Cambridge, UK, 2012.
(29) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci,
B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H.
P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.;
Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima,
T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A.;
Peralta, J. E., Jr.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.;
Kudin, K. N.; Staroverov, V. N.; Keith, T.; Kobayashi, R.; Normand, J.;
Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.;
Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.;
Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.;
Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.;
Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador,
P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.;
Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09
(Revision B.01); Gaussian, Inc.: Wallingford, CT, 2010.
(14) Nakagawa-Goto, K.; Chang, P. C.; Lai, C. Y.; Hung, H. Y.;
Chen, T. H.; Wu, P. C.; Zhu, H.; Sedykh, A.; Bastow, K. F.; Lee, K. H.
Antitumor agents. 280. Multidrug resistance-selective desmosdumotin
B analogues. J. Med. Chem. 2010, 53, 6699−6705.
(15) Nakagawa-Goto, K.; Wu, P. C.; Lai, C. Y.; Hamel, E.; Zhu, H.;
Zhang, L.; Kozaka, T.; Ohkoshi, E.; Goto, M.; Bastow, K. F.; Lee, K.
H. Antitumor agents. 284. New desmosdumotin B analogues with
bicyclic B-ring as cytotoxic and antitubulin agents. J. Med. Chem. 2011,
54, 1244−1255.
(16) Kelder, J.; Grootenhuis, P. D.; Bayada, D. M.; Delbressine, L. P.;
Ploemen, J. P. Polar molecular surface as a dominating determinant for
oral absorption and brain penetration of drugs. Pharm. Res. 1999, 16,
1514−1519.
(17) Egan, W. J.; Merz, K. M., Jr.; Baldwin, J. J. Prediction of drug
absorption using multivariate statistics. J. Med. Chem. 2000, 43, 3867−
3877.
(18) Clark, D. E. Rapid calculation of polar molecular surface area
and its application to the prediction of transport phenomena. 1.
Prediction of intestinal absorption. J. Pharm. Sci. 1999, 8, 807−814.
(19) cLogP and PSA were calculated by using ChemBioDraw 13
software, as described in the Supporting Information.
(20) Matsunaga, N.; Kaku, T.; Itoh, F.; Tanaka, T.; Hara, T.; Miki,
H.; Iwasaki, M.; Aono, T.; Yamaoka, M.; Kusaka, M.; Tasaka, A.
C17,20-Lyase inhibitors. I. Structure-based de novo design and SAR
study of C17,20-lyase inhibitors. Bioorg. Med. Chem. 2004, 12, 2251−
2273.
(21) Campaigne, E.; Bosin, T.; Neiss, E. S. The sulfur analog of
serotonin. J. Med. Chem. 1967, 10, 270−271.
(22) Fournier, J.; Chabert, D.; Joucla, L.; David, E.; Lemaire, M. An
efficient phosphine-free palladium coupling for the synthesis of new 2-
arylbenzo[b]thiophenes. Tetrahedron 2004, 60, 3221−3230.
(23) Ravelli, R. B.; Gigant, B.; Curmi, P. A.; Jourdain, I.; Lachkar, S.;
Sobel, A.; Knossow, M. Insight into tubulin regulation from a complex
with colchicine and a stathmin-like domain. Nature 2004, 428, 198−
202 (see also the Supporting Information).
(24) Dorleans, A.; Gigant, B.; Ravelli, R. B.; Mailliet, P.; Mikol, V.;
Knossow, M. Variations in the colchicine-binding domain provide
insight into the structural switch of tubulin. Proc. Natl. Acad. Sci. U. S.
A. 2009, 106, 13775−13779.
(25) Hamel, E. Evaluation of antimitotic agents by quantitative
comparisons of their effects on the polymerization of purified tubulin.
Cell Biochem. Biophys. 2003, 38, 1−22.
́
(26) Verdier-Pinard, P.; Lai, J. Y.; Yoo, H. D.; Yu, J.; Marquez, B.;
Nagle, D. G.; Nambu, M.; White, J. D.; Falck, J. R.; Gerwick, W. H.;
2389
DOI: 10.1021/jm501859j
J. Med. Chem. 2015, 58, 2378−2389