A Series of Enthalpically Optimized Docetaxel Analogues Exhibiting Enhanced Antitumor Activity and Water Solubility

Article abstract of DOI:10.1021/acs.jnatprod.7b00857

A dual-purpose strategy aimed at enhancing the binding affinity for microtubules and improving the water solubility of docetaxel led to the design and synthesis of a series of C-2- and C-3′-modified analogues. Both aims were realized when the C-3′ phenyl group present in docetaxel was replaced with a propargyl alcohol. The resulting compound, 3f, was able to overcome drug resistance in cultured P-gp-overexpressing tumor cells and showed greater activity than docetaxel against drug-resistant A2780/AD ovarian cancer xenografts in mice. In addition, the considerably lower hydrophobicity of 3f relative to both docetaxel and paclitaxel led to better aqueous solubility. A molecular model of tubulin-bound 3f revealed novel hydrogen-bonding interactions between the propargyl alcohol and the polar environment provided by the side chains of Ser236, Glu27, and Arg320.

Full text of DOI:10.1021/acs.jnatprod.7b00857

Article
pubs.acs.org/jnp
Cite This: J. Nat. Prod. XXXX, XXX, XXX−XXX
A Series of Enthalpically Optimized Docetaxel Analogues Exhibiting
Enhanced Antitumor Activity and Water Solubility
Yun-Tao Ma,† Yanting Yang,‡ Pei Cai,† De-Yang Sun,† Pedro A. Sanchez-Murcia,§ Xiao-Ying Zhang,†
́
Wen-Qiang Jia,† Lei Lei,‡ Mengqi Guo,‡ Federico Gago,§ Hongbo Wang,*,‡ and Wei-Shuo Fang*,†
†State Key Laboratory of Bioactive Substances and Functions of Natural Medicines, Institute of Materia Medica, Chinese Academy of
Medical Sciences and Peking Union Medical College, Beijing 100050, People’s Republic of China
‡Key Laboratory of Molecular Pharmacology and Drug Evaluation (Yantai University), Ministry of Education, Collaborative
Innovation Center of Advanced Drug Delivery System and Biotech Drugs in Universities of Shandong, Yantai University, Yantai
264005, People’s Republic of China
§
́
Area de Farmacología, Departamento de Ciencias Biomed
́ ́
icas, Unidad Asociada al Instituto de Química Medica del CSIC,
Universidad de Alcala,
́
E-28805 Alcala de Henares, Madrid, Spain
́
S
* Supporting Information
ABSTRACT: A dual-purpose strategy aimed at enhancing the
binding affinity for microtubules and improving the water
solubility of docetaxel led to the design and synthesis of a
series of C-2- and C-3′-modified analogues. Both aims were
realized when the C-3′ phenyl group present in docetaxel was
replaced with a propargyl alcohol. The resulting compound, 3f,
was able to overcome drug resistance in cultured P-gp-
overexpressing tumor cells and showed greater activity than
docetaxel against drug-resistant A2780/AD ovarian cancer
xenografts in mice. In addition, the considerably lower
hydrophobicity of 3f relative to both docetaxel and paclitaxel led to better aqueous solubility. A molecular model of tubulin-
bound 3f revealed novel hydrogen-bonding interactions between the propargyl alcohol and the polar environment provided by
the side chains of Ser236, Glu27, and Arg320.
ancer is the second leading cause of death globally,
C
according to epidemiological data from WHO, and was
responsible for 8.8 million deaths in 2015.1 Numerous efforts
on the development of targeted therapy have been made in the
field of anticancer drug research in recent decades to fight this
widespread disease. However, traditional cytotoxic drugs still
play a very important role in cancer therapy. These agents can
be used singly or combined as well as in combination with the
so-called molecular-targeted antitumor agents. Such cytotoxic
drugs include paclitaxel (1a, Figure 1) and its semisynthetic
analogue docetaxel (1b, Figure 1), two tubulin-binding natural-
product-derived molecules,2 which have had documented great
success in the treatment of ovarian, breast, lung, and various
other cancers3−5 since they were launched to market in the
1990s. Despite the widespread use of taxane-based antitubulin
agents, both their compromised activity toward multidrug-
resistant tumors and poor water solubility (causing formulation
and pharmacokinetic problems) restrict their clinical uses.6,7
Investigations have revealed numerous antitumor drug-
resistance mechanisms for taxane-based antitubulin drugs, of
which three are important or clinically relevant:8,9 (a)
overexpression of ATP-dependent drug transporters, for
example, P-glycoprotein (P-gp); (b) overexpression of tubulin
isotypes with reduced drug-binding affinities, mainly βIII
tubulin; and (c) mutations in the taxane-binding site.
Figure 1. Structures of paclitaxel (1a), docetaxel (1b), cabazitaxel
(1c), CTX-63 (1d), LX-2-32C (1e), and Yg-3-46a (1f).
To overcome these drawbacks, modified taxanes with better
antitumor activities and pharmacokinetic (PK) properties have
been pursued.10,11 However, none of them reached the market
Special Issue: Special Issue in Honor of Susan Horwitz
Received: October 10, 2017
© XXXX American Chemical Society and
American Society of Pharmacognosy
A
DOI: 10.1021/acs.jnatprod.7b00857
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Scheme 1. Synthesis of Taxanes 2a−2g
Table 1. Microtubule Binding Constants and Cytotoxicities of Taxanes 2 and Reference Drugs
IC50 (nM)
a
b
c
d
compound
Kb 35 °C (M−1
)
A2780
A2780AD
R/S
HeLa
HeLa/βIII
R/S
paclitaxel (1a)
3.3 × 106
2.1 × 107
6.1 × 108
2.4 × 107
2.8 × 107
2.7 × 107
6.1 × 106
6.4 × 106
4.7 × 106
13.4
11.2
2.6
16040
21650
1171
144.0
1750
376.0
1947
7561
2227
1197
1933
450
5.0
82
4.6
3.5
1.6
2.9
4.2
4.3
7.5
45.4
21.7
41.0
27.1
9.4
8.9
docetaxel (1b)
7.7
2a
2b
2c
2d
2e
2f
5.9
28.6
21.3
21.6
31.9
153.0
38.8
43.2
47.5
108.0
461.0
242.0
13.4
10.3
11.0
14.4
10.2
11.0
17
61
49
2g
30.8
72
a
b
Resistant ovarian carcinoma cells overexpressing P-gp. The relative resistance values were calculated by dividing the IC50 value of the resistant
c
ovarian carcinoma cells by the IC50 value of the nonresistant ovarian carcinoma cells. βIII transfected human cervical carcinoma cells HeLa cell line
(wild-type βIII). The relative resistance values were calculated by dividing the IC50 value of the βIII transfected cell line (wild-type βIII) by the IC50
d
value of the parental HeLa cell line.
until cabazitaxel (1c, Figure 1), a poor substrate of P-gp that
remains active in docetaxel-resistant tumors,12 was approved by
the U.S. FDA in 2010 for the treatment of metastatic refractory
prostate cancer.13
equally or more hydrophobic than this reference taxane. As it
has been recognized that the “molecular obesity”19,20 during the
lead optimization process may result in poor PK properties, the
discovery of 1d through setting up a new water-shielded
hydrogen bond between a hydroxy group in the ligand and
tubulin, i.e., an enthalpy-driven process, could be valuable in
finding new drug candidates with better PK and also improved
water solubility.
In this study, further optimization of 1d was effected by
modifications at C-2 and/or C-3′ of this taxane, as both
positions were known to greatly impact taxane−tubulin
interactions and also cytotoxic activity.11,17,21,22 It was hoped
to optimize the semisynthetic taxanes through an enthalpically
driven process, thus retaining and/or improving pharmacody-
namics (PD) together with PK properties. As a result of these
efforts, a C-3′-modified docetaxel analogue, 3f, was found that
exhibits more potent antitumor activity than docetaxel in vivo.
In the earlier study,14 it was found that enhancing the
microtubule (MT) binding affinity of taxanes could be utilized
to combat drug-resistant tumors overexpressing P-gp. This new
strategy was also confirmed in subsequent research leading to
CTX-63 (1d, Figure 1), a reported C-3′-modified docetaxel
analogue that was successfully predicted as a tighter MT-
binding molecule in the comparative binding energy
(COMBINE) analysis. The higher affinity of 1d for β-tubulin
resulted from a more favorable interaction between the benzylic
hydroxy group at C-3′ and the side chain of Glu27 in the
paclitaxel-binding site.15
Previously, two paclitaxel analogues, LX-2-32C (1e, Figure
1)14,16 and Yg-3-46a (1f, Figure 1),17,18 were selected as
candidates for antitumor drug development, both of which are
more potent than paclitaxel in drug-resistant tumors and also
B
DOI: 10.1021/acs.jnatprod.7b00857
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
RESULTS AND DISCUSSION
■
C-2-Modified Docetaxel Analogues. C-2 substitutions
are known to be critical for the taxane−MT binding affinity.17
Several optimal substituents were chosen from literature
data23,24 to replace the 2-benzoate in CTX-63 (1d). The
synthesis of the C-2-modified CTX-63 was similar to that in a
previous study,15 but was optimized by replacement of a 2′-O-
silyl protective group to improve the overall yield. Coupling of
the protected baccatin III 15 and the corresponding β-lactam
(Scheme S1, Supporting Information) furnished the inter-
mediate taxane 16 bearing a C-3′ m-TBS-O-hydroxymethyl-
phenyl group, which was subjected to hydrolysis, esterification,
and deprotection, successively. More specifically, treatment of
16 with anhydrous potassium hydroxide afforded the C-2
debenzoyl intermediate 17 in 41% yield based on 44% of
substrate 16 recovery. Taxanes 2a−2f were afforded via the
esterification of 2-debenzoyl intermediate 17 with various
substituted benzoic acids in the presence of dicyclohexylcarbo-
diimide (DCC) and 4-dimethylaminopyridine (DMAP),
followed by desilyation, while 2g was obtained by the
deprotection of 16 (Scheme 1). It is worth mentioning that
the tert-butyldimethylsilyl (TBS) group was chosen to replace a
triethylsilyl (TES) group in the previous report,15 as the TES
group was removed during the C-2 esterification step, and C-2
and C-2′ diesterified products were obtained almost exclusively.
C-2 derivatives of 1d were then tested in both molecular and
cellular assays (Table 1). The highest MT binding affinity
derivative 2a found was the 2-m-azido-substituted compound,
as in most taxane studies. Compound 2a was also the most
cytotoxic compound in two pairs of drug-sensitive and
-resistant (P-gp and β-III tubulin overexpression mechanisms)
tumor cells among the series 2a−2g, with a 29-fold enhance-
ment in MT binding affinity, and was 4−5 times more cytotoxic
against A2780 drug-sensitive cells and 18−19 times more active
against A2780AD-resistant cells (P-gp overexpressed). This,
again, demonstrated the utility of the proposed strategy.14,17
None of the derivatives other than 2a was better than the
reference compounds 1a and 1b for both HeLa and β-III
transfected drug-resistant HeLa/βIII cells. However, 2a did not
exhibit superior activity to docetaxel in A2780 and A2780AD
xenografted mice. This fact significantly reduces the possibility
of progressing the C-2-modified CTX-63 as a new drug
candidate.
Figure 2. Structures of the C-3′-modified CTX-63 analogues 3a−3i.
deprotection formed aldehyde 3a. Taxane 3b was obtained by
methylation in the presence of di-tert-butyl-4-methylpyridine
(DTBMP) and methyl iodide and deprotection. N-Methyl-
morpholine-N-oxide (NMO) oxidation of 22 catalyzed with
tetrapropylammonium perruthenate (TPAP) furnished a m-
formyloxy-substituted compound, which was deprotected to
afford 3c. Finally, 3d was synthesized by following a three-step
procedure, namely, attachment of a leaving O-tosyl (OTs)
group onto the m-OH on C-3′ phenyl, nucleophilic
replacement of OTs with dimethylamine, and deprotection.
The β-lactam 35a bearing a C-3′ m-(2-hydroxyethyl)phenyl
substitution was synthesized following a route similar to that
leading to β-lactam 13 (Scheme 3). Treatment of 2-(3-
bromophenyl)ethanol 23 with n-butyllithium and dimethylfor-
mamide (DMF) afforded aldehyde 24, which was then
protected by a TBS group. A Staudinger reaction was applied
to the adduct formation of imine 30a and ketene prepared from
acetoxyacetyl chloride to furnish racemic β-lactam 31a. During
removal of the p-methoxylphenyl (PMP) group by ceric
ammonium nitrate (CAN) oxidation, TBS was also removed to
afford racemic lactam 32a. After kinetic resolution of racemate
32a, protective groups were reintroduced to the lactam 33a,
affording β-lactam 35a over three steps. Coupling of 7,10-O-
di(triethylsilyl)-10-DAB (15) with β-lactam followed by
deprotection afforded 3e with one more carbon having meta-
substitution on the C-3′ phenyl. The extension of the meta-
alkyl chain on phenyl aimed to optimize the interaction of the
ligand with the surroundings of Glu 27 in β-tubulin, as revealed
in the previously reported COMBINE model.15
C-3′-Modified Docetaxel Analogues. Next, C-3′ of 1d
was explored through the introduction of various hydrogen
bond (HB) donors and acceptors, as well as introducing
negatively and positively charged groups, into this position. The
rigidity of the linker (phenyl ring) between C-3′ and the
terminal functional group (hydroxy) was explored by replacing
the phenyl with alkynyl, alkenyl, and alkanyl moieties.
In general, the corresponding β-lactams were first synthe-
sized and then coupled to protected baccatin III, while during
the preparation of 3a−3d functional group interconversion was
applied, thereby converting 22 (protected 1d) into 3a−3d.
A routine β-lactam synthetic route15 furnished 20, in which
the OH groups were protected by TBS and TES. The β-lactam
20 was coupled to 7,10-O-di(triethylsilyl)-10-deacetylbaccatin
III (DAB) (15) to obtain 21. Selectively removing the TES
protective group on benzyl alcohol by HF/pyridine (Py) in an
ice bath afforded the intermediate 22, which was then
converted to taxanes 3a−3d under various conditions (Scheme
2). Treatment of 22 with pyridinium dichromate (PDC) and
Further optimization of C-3′ substituents was performed by
replacing the phenyl with alkynyl, alkenyl, and alkanyl moieties.
The alkynyl-substituted β-lactam 35b was synthesized from
2-butyne-1,4-diol (26) as the starting material (Scheme 3).
Selective monohydroxyl protection of 26 followed by oxidation
afforded 25b, which was transformed into imine 30b and then
formed β-lactam adduct 31b. After several steps of trans-
formation including resolution, chiral β-lactam 35b was
obtained and coupled with 7,10-O-di(triethylsilyl)-10-DAB
(15) followed by deprotection to afford taxane 3f. Upon
hydrogenation of 3f with Lindar’s catalyst or Pd/C, cis-alkenyl-
bearing 3g and alkanyl 3i were obtained.
However, the trans-alkenyl analogue 3h could not be
obtained directly by reduction of 3f, so β-lactam 35c was
synthesized alternatively, that is, starting from (E)-but-2-ene-
1,4-diol 28 (Scheme 3) and following a procedure similar to
that leading to the synthesis of β-lactam 35b. Coupling of 35c
with 7,10-O-di(triethylsilyl)-10-DAB (15) followed by depro-
tections afforded 3h.
C
DOI: 10.1021/acs.jnatprod.7b00857
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Scheme 2. Synthesis of Taxanes 3a−3d
Among the C-3′-modified taxanes obtained, only 3f and 3h
showed higher binding affinities for tubulin relative to that of
docetaxel, with 3f bearing a 3′-(3-hydroxylprop-1-ynyl) group
being endowed with ca. 8 times higher affinity. The extension
of one carbon into m-hydroxymethylphenyl or the replacement
of the phenyl with a cis-/trans-ethylene or an ethyl group
resulted in as much as a 500-fold reduced binding affinity.
Interestingly, C-3′-modified taxanes bearing hydroxyformyl
(3c) and dimethylamino (3d) groups showed diminished
binding affinities.
(Figure 4). This proposed binding mode also accounts for the
importance of the spatial position of the hydroxy group in
taxanes 3g−3i, as the E-configuration taxane 3h displayed ca.
146 times higher binding affinity than the Z-configuration
taxane 3g. The lower binding affinity for taxane 3i relative to 3f
could be explained by the much larger flexibility of its side
chain, leading to decreased enthalpic contributions and a higher
entropic penalty.
Next, the cytotoxicities in two pairs of drug-sensitive and
drug-resistant tumor cell lines were assessed (Table 3). The
highest binding-affinity taxane 3f was the most active in the C-
3′-modified series 3, but still less active than paclitaxel (1a) and
docetaxel (1b) in both sensitive and resistant tumor cell lines,
while the resistant indices were improved slightly. Compared
with 2g (the 10-deacetyl derivative of CTX-63), both functional
group replacements (3a and 3b) and prolonged linkage (3e)
decreased the cytotoxicity 1.3- to 109-fold, and introduction of
charged groups (3c and 3d) resulted in complete activity loss.
Taxanes 3a and 3h were similar to or a little less active than 3f
against the two cell lines, and 3g and 3i were much less active.
As the high binding-affinity taxane 3f displayed potent
cytotoxicity against both drug-sensitive tumor cells and drug-
resistant tumor cells, it was selected for animal model
evaluation. In drug-sensitive ovarian cancer A2780 xenografted
mice, 3f showed similar inhibitory effects at different doses (10
and 20 mg/kg) to that of paclitaxel (20 mg/kg) (Table S1,
Supporting Information) and demonstrated superior activity to
docetaxel in mice bearing drug-resistant ovarian cancer A2780/
AD xenografts in a dose-dependent manner (Table S2,
Supporting Information) (Figure 5).
The propargyl alcohol present in taxane 3f is directed, as
reported for the benzyl alcohol of CTX-63,15 toward the
hydroxy group of Ser236 and the carboxylate of Glu27, which is
held in place by the guanidinium groups of Arg243 and Arg320
(Figure 3). During the molecular dynamics simulation,
additional long-lived hydrogen bonds are established between
the oxetane oxygen and the NH of Thr276, the carbonyl O-9
and the side-chain carboxamide of Gln281, the C-3′ benzamide
carbonyl oxygen and Nε of His229, the hydroxy O-2′ and the
backbone carbonyl of Arg369, and, through an intervening
long-residence water molecule, between N-3′ and the
carboxylate of Asp27. In this respect, it was noted that the
impact on binding affinity of the first hydration shell of
protein−ligand complexes is often overlooked, in agreement
with other authors.25 It is believed that the polar environment
surrounding the propargyl alcohol in the bound conformation
enthalpically boosts the binding energy relative to docetaxel
(1b), which has an unsubstituted phenyl ring in this position. In
fact, when calculating the binding energies and component
contributions for the association of 3f and docetaxel with β-
tubulin at the interprotofilament site (Table 2), similar values
within experimental error were obtained. According to these
data, this similarity is most likely due to the fact that the
intended gain in binding enthalpy was obtained through
improved electrostatic interactions and a lower penalty for
receptor desolvation but at the cost of a loss of van der Waal’s
interactions, because of the replacement of the bulky phenyl
ring with the linear alkynyl group. In fact, similar binding
signatures were obtained when pairwise comparing the solvent-
corrected residue-based contributions for these two molecules
To facilitate the better understanding of pharmacodynamics
for 3f, its pharmacokinetics was also analyzed as shown in
Figure 6 and Table S3, Supporting Information. The plasma
drug concentration observed was higher in the 3f liposome
group than that in the docetaxel group at 12, 24, 48, 72, and 96
h postadministration, and the accumulation and clearance of
docetaxel (1b) occurred over a considerably shorter period
when compared with the 3f liposome group.
Water Solubility Evaluations. The initial design for these
CTX-63 analogues also aimed to enhance the water solubility of
D
DOI: 10.1021/acs.jnatprod.7b00857
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Scheme 3. Synthesis of Taxanes 3e−3i
hydrophobic taxanes so as to improve the PK properties as well
as facilitate its formulation. Following a reported procedure,27
the water solubility of the synthesized compounds 2 and 3 was
measured by reversed-phase HPLC (Table 4). All compounds
showed improved water solubility compared to docetaxel and
paclitaxel (31−135 times), among which the most potent
taxane, 3f, was found to be about 3-fold more water soluble
than docetaxel.
As it is known that high-affinity taxanes may overcome P-gp
overexpression-mediated tumor drug resistance,17 semisyn-
thetic taxanes [e.g., Lx-2-32c (1e),14,16 Yg-3-46a (1f)18] with
such properties at the molecular level were prepared. Although
potent cytotoxicity especially against drug-resistant cells was
well established, such highly hydrophobic molecules did
encounter some problems (e.g., unfavorable PK and difficulty
in formulation) in their evaluation.
It was reasoned that such problems are due, at least in part,
to the hydrophobic nature of these semisynthetic taxanes,
which were obtained during the entropy-driven structural
optimization. Hence, an enthalpy-driven optimization process
was preferred, by introducing a hydroxy group onto the C-3′-
phenyl, as shown in a previous report.15 In the present work,
this strategy was extended to the design and synthesis of a
series of more polar, high-affinity taxanes, including the
representative 3f.
Such taxanes (3), while leveraging the advantages against
tumor drug resistance they could offer, are also more water-
soluble than docetaxel (1b). It has been known29 that the ease
of formulation and lack of hypersensitive adverse effect of 1b
on injection (vs 1a) can be attributed to the slightly improved
aqueous solubility of 1b relative to 1a. Thus, it may be expected
that taxane 3f with both favored properties might exhibit better
E
DOI: 10.1021/acs.jnatprod.7b00857
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Figure 4. Calculated solvent-corrected interaction energies (kcal·
mol−1) between individual β-tubulin residues and either 3f (hatched
bars) or docetaxel (empty bars) bound at the interprotofilament site.
For clarity, only the contributions achieving ≤1 kcal·mol−1 are shown,
which together represent a “binding fingerprint”.
Figure 3. Close-up view of a representative snapshot from the
molecular-dynamics (MD) trajectory showing 3f (yellow sticks)
bound at the interprotofilament taxane-binding site of β-tubulin
(blue cartoon). Relevant protein residues are labeled and shown as
sticks. Intermolecular hydrogen bonds are displayed as dashed yellow
lines.
Reagents were purchased from J&K and Alfa Aesar Chemical
Companies. All anhydrous solvents were purified and dried according
to standard procedures, unless otherwise indicated. Reactions were
monitored by TLC (silica gel, GF254) with UV light and H2SO4-
anisaldehyde spray visualization. The purity of the final taxoids was
analyzed by HPLC.
The general procedure for the synthesis of compounds 2, 3, 5−13,
15−22, 24, 25, 27, and 29−35 is provided in the Supporting
Information, where full details of their purification and spectroscopic
data are provided. Such information is given for compound 3f below
with the procedure for the biological testing and molecular modeling.
3′-Dephenyl-3′-(3-hydroxyprop-1-ynyl)docetaxel (3f). To a
stirred solution of 15 (25 mg, 0.0323 mmol) in anhydrous
tetrahydrofuran (THF) (0.42 mL) under argon and cooled to −45
°C was added dropwise lithium hexamethyldisilamide (LHMDS) (1.0
M in THF/ethylbenzene, 48.5 μL, 0.048 mmol). The reaction mixture
was stirred for 20 min at −45 °C, and then the solution of 35b (18 mg,
0.039 mmol) in anhydrous THF (0.1 mL) was added dropwise, the
reaction mixture was stirred for 80 min at the same temperature, and
LHMDS (1.0 M in THF/ethylbenzene, 48.5 μL, 0.048 mmol) was
added dropwise. The reaction mixture was stirred for a further 1 h and
then quenched with saturated aqueous NH4Cl solution (10 mL) and
extracted with ethyl acetate (20 mL). The organic layer was washed
with saturated aqueous NH4Cl solution (10 mL) and brine (10 mL)
and dried over Na2SO4, and solvent was removed under reduced
pressure to obtain the crude product without any further purification.
Then, to a stirred solution of the crude product in acetonitrile (1.4
mL) was added Py/HF (2:1, 1.3 mL, 9.69 mmol), and the reaction
mixture was stirred at room temperature for 24 h. Next, the mixture
was diluted with ethyl acetate (20 mL), washed with saturated aqueous
NaHCO3 solution (10 mL) and brine (10 mL), and dried over
anhydrous Na2SO4. The organic layer was evaporated under reduced
pressure. Purification of the crude product by silica gel chromatog-
raphy (MeOH/CH2Cl2, 1:25) gave 53% yield of 3f as a colorless oil:
1H NMR (CDCl3, 500 MHz) δ 8.10 (2H, d, J = 7.5 Hz, Ar), 7.61 (1H,
t, J = 7.0 Hz, Ar), 7.50 (2H, t, J = 8.0 Hz, Ar), 6.21 (1H, t, J = 8.5 Hz,
H-13), 5.67 (1H, d, J = 6.5 Hz, H-2), 5.36 (1H, d, J = 6.5 Hz, H-3′),
5.25 (1H, s, H-10), 4.99−4.95 (2H, m, H-5, NH), 4.44 (1H, s, H-2′),
4.33−4.18 (5H, m, H-20, H-7, CH2), 3.91 (1H, d, J = 7.0 Hz, H-3),
2.63−2.55 (1H, m, H-6a), 2.41 (3H, s, OAc), 2.32−2.25 (2H, m, H-
14), 1.95 (3H, s, CH3), 1.89−1.84 (1H, m, H-6b), 1.76 (3H, s, CH3),
antitumor activity, especially against drug-resistant tumors. In
xenograft animal testing, the antitumor effect of 3f was
comparable to those of the reference taxanes in drug-sensitive
A2780 and much better in drug-resistant A2780AD xenografts.
Such superior in vivo activity may be explained by not only
slightly improved tubulin binding and thus cytotoxicity but also
enhanced drug exposure (larger AUC and Cmax, Table S3,
Supporting Information) and prolonged elimination (lower
clearance, Table S3, Supporting Information) in animal models.
To assess the drug-like properties across the semisynthetic
taxane family, the logP and ligand efficiency (LE) for
commercial drugs 1a and 1b, hydrophobic taxanes 1e and 1f
investigated as drug candidates, and two high-affinity taxanes
(2a and 3f) from this study were calculated. It was found that
the LEs are comparable for all compounds, and the much
higher clogPs for analogues 1e and 1f relative to marketed
drugs may be related to their poor drug-like properties. In
contrast, the lower clogP value for taxane 3f is suggestive of an
improved pharmaceutical profile (Table S4, Supporting
Information).
In summary, enthalpy-driven optimization led to the
discovery of semisynthetic taxanes inclusive of 3f, which
demonstrated improved pharmaceutical properties over
commercially used taxane and thus the potential to be an
antitumor drug.
EXPERIMENTAL SECTION
■
1
General Experimental Procedures. H and 13C NMR spectra
were recorded on Varian 300, 400, and 500 MHz NMR spectrometers
or a Bruker AVANCE III 600 MHz NMR spectrometer. Mass spectra
(ESI) were measured on a JEOL Accu TOF CS (JMS T100CS).
a
Table 2. Binding Energies (kcal·mol−1) and Component Contributions26 for the Association of 3f and Docetaxel with β-
Tubulin at the Interprotofilament Site
complex
total
van der Waal’s
coulombic
ligand desolvation
receptor desolvation
apolar
3f
docetaxel (1b)
−65.3 4.5
−67.4 5.9
−66.0 3.4
−70.6 3.7
−8.6 1.9
−7.1 1.7
4.9 0.8
4.6 0.7
10.2 2.5
11.6 2.7
−5.8 0.2
−6.0 0.2
a
Mean values and standard deviations calculated from 250 snapshots taken from the MD trajectories (25 ns).
F
DOI: 10.1021/acs.jnatprod.7b00857
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Table 3. Microtubule Binding Constants and Cytotoxicities of Taxanes 3 and Reference Drugs
IC50 (nM)
a
b
c
d
compound
Kb 35 °C (M−1
)
A2780
A2780AD
R/S
HeLa
HeLa/βIII
R/S
paclitaxel (1a)
3.3 × 106
2.1 × 107
6.2 × 106
3.5 × 105
<2 × 104
<2 × 104
3.5 × 105
1.8 × 108
3.5 × 105
5.1 × 107
9.8 × 106
13.4
16 040
21 650
26 000
24 680
15 650
ND
1197
1933
1070
272
1.8
4.6
41.0
8.9
7.7
21.3
5.4
0.2
7.1
3.4
6.0
2.1
7.6
7.7
docetaxel (1b)
11.2
3.5
27.1
3a
3b
3c
3d
3e
3f
24.3
28.8
612.5
720.0
3077
25 050
7961
191.6
1466
404.6
809.0
90.6
134.1
14 140
3527
2356
32.2
8934
>10 000
1926
25.8
ND
9.8
18 830
18 770
20 630
25 310
21 250
728
44.6
343
280
3g
3h
3i
462.7
73.8
707.3
53.3
75.8
104.4
a
b
Resistant ovarian carcinoma cells overexpressing P-gp. The relative resistance values were calculated by dividing the IC50 value of the resistant
c
ovarian carcinoma cells by the IC50 value of the nonresistant ovarian carcinoma cells. βIII transfected human cervical carcinoma cells HeLa cell line
(wild-type βIII). The relative resistance values were calculated by dividing the IC50 value of the βIII transfected cell line (wild-type βIII) by the IC50
d
value of the parental HeLa cell line.
Figure 5. Inhibitory effects of 3f on the tumor growth of A2780 and A2780AD in nude mice.
1.36 (9H, s, t-BuO), 1.21 (3H, s, CH3), 1.13 (3H, s, CH3); 13C NMR
(CDCl3, 150 MHz) δ 211.2, 171.2, 170.6, 167.0, 154.9, 138.4, 136.0,
133.7, 130.1, 129.1, 128.7, 84.1, 83.1, 81.4, 81.1, 78.8, 76.6, 74.8, 74.6,
72.4, 72.0, 60.4, 57.7, 46.6, 43.1, 37.0, 35.6, 28.2, 26.4, 22.5, 20.5, 14.2,
9.9; ESIMS m/z 786.3 [M + H]+, 808.3 [M + Na]+.
In Silico Model Building, Molecular Simulations and Binding
Energy Analysis. The simulated macromolecular ensemble repre-
senting a short piece of a microtubule with bound docetaxel (DXL,
1b) and 3f was built as previously reported for dictyostatin.30 Briefly,
(i) α-subunits A, E, I, and K and β-subunits B and F were selected
from the cryo-electron microscopy reconstructions (PDB code 3J6G)
of kinesin-decorated MTs in complex with paclitaxel solved at ∼5 Å
resolution;31 (ii) missing residues 39−48 in the four α-subunits,
together with the partially hydrated Ca2+ ion coordinated by Asp39,
Thr41, Gly44, and Glu55 were added; (iii) the two guanosine
diphosphate (GDP) and four guanosine triphosphate (GTP)
molecules were present in the nucleotide-binding sites; and (iv) the
docked poses of DXL and CTX-63 within the taxane-binding site in an
α,β-tubulin dimer reported previously15 were used to produce two
different complexes in which DXL and 3f occupied either the
interprotofilament or the solvent-exposed site in two GDP and four
GTP molecules’ formation could be separately assessed.
Geometry optimization of DXL and 3f was achieved by means of
the updated AM1 Hamiltonian,32 as implemented in the sqm
program,33 which also produced the atomic charge distributions for
both ligands. The addition of missing hydrogen atoms to the protein
G
DOI: 10.1021/acs.jnatprod.7b00857
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
modified as previously described.14 Cell cycle analysis was performed
as previously described.45
Preparation of Liposome 3f. Liposomes 3f were prepared using
a film dispersion method following the published protocol.46 Briefly,
3f, lecithin, and cholesterol were dissolved initially in chloroform, and
then the organic solvent was evaporated to obtain a membrane after 5
min of stirring. The resulting membrane was dissolved by the addition
of phosphate-buffered saline to obtain a liposome solution. The
solution obtained using this process was disrupted in an ultrasonic
homogenizer. Then, the remaining solution was lyophilized to dried
19c liposomes using a freeze-dryer system (Labconco), and their
microstructure was observed using a scanning electron microscope.
Particle size was evaluated by means of a particle size analyzer
(Mastersizer 2000, Malvern Instruments), and the encapsulation
efficacy was measured using high-performance liquid chromatography
(HPLC; Shimadzu LC-20A, C18 column).
Xenograft Studies. Nude mice (6−8 weeks old, BALB/c, female)
were used to establish the xenograft tumors following the published
protocol.47 Briefly, A2780/AD cells (1 × 107) or A2780 cells (5 × 106)
were implanted in the dorsal region of recipient mice by means of
subcutaneous injection. Once a tumor had reached around 300 mm3 in
size, the mice were randomized into five groups as control, paclitaxel
(20 mg/kg), docetaxel (10 mg/kg), 3f-low (10 mg/kg), and 3f-high
(20 mg/kg) with six or seven mice per group. The animals were
administrated by intraperitoneal injection twice a week. Tumor growth
and body weight were measured every three days during the treatment.
At the end of the treatment, mice were sacrificed and tumors were
removed and weighed. The use of animals was approved by the Animal
Experimentation Ethics Committee of Yantai University (protocol
number 20170605) in accordance with the guidelines for ethical
conduct in the care and use of animals.
Pharmacokinetic Experiments. The pharmacokinetic profile of
the 3f in liposomes and Taxotere (docetaxel injection) were explored
using SD rats (protocol number 20170711), after a single intra-
peritoneal injection at a dose of 10 mg/kg. Blood samples of about 500
μL were collected via the posterior orbit into a heparinized vacutainer
tube at 0, 0.17, 0.25, 0.5, 1, 2, 4, 8, 12, 24, 48, 72, and 96 h
postadministration. The blood samples were centrifuged at 4000 rpm
for 10 min at 4 °C to obtain plasma samples, which were kept frozen at
−80 °C until analysis. Each plasma sample (50 μL) was treated with
200 μL of acetonitrile and vortexed for 30 s followed by centrifugation
at 12 000 rpm for 10 min at 4 °C. Then, the supernatants were
transferred to other test tubes. The 3f concentration in the plasma
samples was quantified using a Diamonsil C18 (4.6 mm × 6250 mm;
⦶ 5 mm) HPLC column at a temperature of 30 °C. The mobile phase
was 55:45 (acetonitrile/water) for 3f and 70:30 (acetonitrile/water)
for docetaxel at a flow rate of 1 mL/min. The effluent was detected at
230 nm, and the area under the peak was used for quantification.
Pharmacokinetic parameters were calculated using the noncompart-
mental analysis with DAS 2.0 (Drug and Statistics) Software.
Partition Coefficient (LogP Value) Calculations. CLogP value
calculations were performed with ChemBioDraw Ultra 11.0 and
Accelrys Discovery Studio 2.5. Structures of ligands were drawn using
ChemBioDraw Ultra 11.0 and saved as sdf file. The sdf file was opened
by Accelrys Discovery Studio 2.5, and cLogP values were calculated by
Protocols-General Purpose-Calculate Molecular Properties as “ALogP”
in the software. The results are shown in Table S4, Supporting
Information.
Figure 6. Docetaxel and 3f plasma concentration−time profile in SD
rats following a single ip dose at 10 mg/kg. Data are expressed as mean
SD (n = 3).
Table 4. Water Solubility of Taxanes 2 and 3
water
water
solubility (mg/
mL)
a
solubility
compound
(mg/mL)
fold compound
fold
31
paclitaxel
(1a)
0.0003
1
3a
0.0093
b
docetaxel
(1b)
0.006−0.007
20−
3b
0.0143
48
23
2a
2b
2c
2d
2e
2f
0.0126
0.0109
0.0111
0.0125
0.0149
0.0112
0.0095
42
36
37
42
50
37
32
3c
3d
3e
3f
0.0406
0.0244
0.0289
0.0192
0.0186
0.0137
0.0274
135
81
96
64
62
46
91
3g
3h
3i
2g
a
Water solubility was determined by reversed-phase HPLC at 25 °C.
Data from ref 28.
b
ensemble and computation of the protonation state of titratable
groups at pH 6.5 were carried out using the H++ 3.0 Web server.34
The ff14SB force field parameter set35 in AMBER 1436 was used to
assign bonded and nonbonded parameters to protein and ligand
atoms. Each complex was immersed in a cubic box containing
∼110 000 TIP3P water molecules plus >100 Na+ ions to achieve
electroneutrality and was simulated under periodic boundary
conditions in the absence of any external restraints for 50 ns at 300
K. Electrostatic interactions were treated using the smooth particle
mesh Ewald method37 with a grid spacing of 1 Å. The cutoff distance
for the nonbonded interactions was 9 Å, and the SHAKE algorithm38
was applied to all bonds involving hydrogens so as to allow an
integration step of 2.0 fs to be used. Subsequent gradual cooling, from
300 K to 273 K over 1 ns, of snapshots taken regularly every 2.5 ns,
followed by energy minimization until the root-mean-square of the
Cartesian elements of the gradient was less than 0.1 kcal·mol−1·Å−1,
provided representative structures for the complexes. These sets of
optimized coordinates were analyzed using the cpptraj routines
implemented in the AmberTools14 suite39 and the in-house MM-
ISMSA software,26 which provided the solvent-corrected binding
energies as well as their decomposition into van der Waal’s,
Coulombic, apolar, and desolvation contributions.
ASSOCIATED CONTENT
■
Binding Affinity Measurement. Calf brain tubulin was purified
as described.40 Glutaraldehyde-stabilized microtubules with active
paclitaxel binding sites were prepared as described.41 The binding
constants of the compounds to these microtubules were measured as
previously described.42
S
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.jnat-
prod.7b00857.
Cell-Based Assays. Cytotoxicity evaluation was performed with
A2780, A2780AD, (overexpressing P-gp),43 human ovarian carcinoma
cell lines, HeLa, and HeLa-βIII transfected cells44 with the MTT assay
Addtional information (PDF)
Spectra (PDF)
H
DOI: 10.1021/acs.jnatprod.7b00857
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
(17) Matesanz, R.; Barasoain, I.; Yang, C.-G.; Wang, L.; Li, X.; Ines,
C.; Coderch, C.; Gago, F.; Jimenez-Barbero, J.; Andreu, J. M.; Fang,
W. S.; Diaz, J. F. Chem. Biol. 2008, 15, 573−585.
(18) Cai, P.; Lu, P.-H.; Sharom, F. J.; Fang, W.-S. Cancer Lett. 2013,
341, 214−223.
(19) Meanwell, N. A. Chem. Res. Toxicol. 2011, 24, 1420−1456.
(20) Hann, M. M. MedChemComm 2011, 2, 349−355.
(21) Ojima, I.; Slater, J. C.; Michaud, E.; Kuduk, S. D.; Bounaud, P.-
Y.; Vrignaud, P.; Bissery, M.-C.; Veith, J. M.; Pera, P.; Bernacki, R. J. J.
Med. Chem. 1996, 39, 3889−3896.
(22) Tang, Y.; Rodriguez-Salarichs, J.; Zhao, Y.; Cai, P.; Estevez-
Gallego, J.; Balaguer-Perez, F.; Horcajo, M. R.; Lucena-Agell, D.;
Barasoain, I.; Diaz, J. F.; Fang, W.-S. Eur. J. Med. Chem. 2017, 137,
488−503.
AUTHOR INFORMATION
Corresponding Authors
*Tel: +86 138 5357 9172. E-mail: hongbowangyt@gmail.com
(H. Wang).
*Tel: +86 010-63165229. Fax: +86 010-63017757. E-mail:
wfang@imm.ac.cn (W.-S. Fang).
■
ORCID
Yun-Tao Ma: 0000-0002-8961-9445
Federico Gago: 0000-0002-3071-4878
Notes
The authors declare no competing financial interest.
(23) Kingston, D. G. I.; Chaudhary, A. G.; Chordia, M. D.; Gharpure,
M.; Gunatilaka, A. A. L.; Higgs, P. I.; Rimoldi, J. M.; Samala, L.; Jagtap,
P. G. J. Med. Chem. 1998, 41, 3715−3726.
ACKNOWLEDGMENTS
■
The authors acknowledge the financial support by NSFC (grant
no. 30930108), CIFMS (grant no. 2016-I2M-1-010), and the
Spanish MINECO (grant no. SAF2015-64629-C2-2-R).
(24) Kingston, D. G. I.; Newman, D. J. Curr. Opin. Drug Discovery
Dev. 2007, 10, 130−144.
(25) Cramer, J.; Krimmer, S. G.; Heine, A.; Klebe, G. J. Med. Chem.
2017, 60, 5791−5799.
DEDICATION
■
́
́
(26) Klett, J.; Nunez-Salgado, A.; Dos Santos, H. G.; Cortes-Cabrera,
̃
Dedicated to Dr. Susan Band Horwitz, of Albert Einstein
College of Medicine, Bronx, NY, for her pioneering work on
bioactive natural products.
́
A.; Perona, A.; Gil-Redondo, R.; Abia, D.; Gago, F.; Morreale, A. J.
Chem. Theory Comput. 2012, 8, 3395−3408.
(27) Ali, S. M.; Hoemann, M. Z.; Aube, J.; Mitscher, L. A.; Georg, G.
I. J. Med. Chem. 1995, 38, 3821−3828.
REFERENCES
(28) Du, W.; Hong, L.; Yao, T.; Yang, X.; He, Q.; Yang, B.; Hu, Y.
Bioorg. Med. Chem. 2007, 15, 6323−6330.
■
(1) http://www.who.int/mediacentre/factsheets/fs297/en/.
(2) Yang, C.-P. H.; Horwitz, S. B. Int. J. Mol. Sci. 2017, 18, 1733−
1743.
(29) Hennenfent, K. L.; Govindan, R. Ann. Oncol. 2006, 17, 735−
749.
(3) Ozols, R. F. Semin. Oncol. 1994, 21, 1−9.
(30) Trigili, C.; Barasoain, I.; San
Redondo-Horcajo, M.; Nogales, A.; Gardner, N. M.; Meyer, A.;
Naylor, G. J.; Gomez-Rubio, E.; Gago, F.; Steinmetz, M. O.; Paterson,
́
chez-Murcia, P. A.; Bargsten, K.;
(4) Seidman, A. D.; Reichman, B. S.; Crown, J. P.; Yao, T. J.; Currie,
V.; Hakes, T. B.; Hudis, C. A.; Gilewski, T. A.; Baselga, J.; Forsythe, P.;
Lepore, J.; Marks, L.; Fain, K.; Souhrada, M.; Onetto, N.; Arbuck, S.;
Norton, L. J. Clin. Oncol. 1995, 13, 1152−1159.
(5) Ojima, I.; Lichtenthal, B.; Lee, S.; Wang, C.-W.; Wang, X. Expert
Opin. Ther. Pat. 2016, 26, 1−20.
́
I.; Prota, A. E.; Díaz, J. F. ACS Omega 2016, 1, 1192−1204.
(31) Alushin, G. M.; Lander, G. C.; Kellogg, E. H.; Zhang, R.; Baker,
D.; Nogales, E. Cell 2014, 157, 1117−1129.
(32) Rocha, G. B.; Freire, R. O.; Simas, A. M.; Stewart, J. J. P. J.
Comput. Chem. 2006, 27, 1101−1111.
(6) Singla, A. K.; Garg, A.; Aggarwal, D. Int. J. Pharm. 2002, 235,
179−192.
(7) Casazza, A. M.; Fairchild, C. R. In Drug Resistance; Hait, W. N.,
Ed.; Kluwer Academic Publishers: Boston, 1996; Vol. 87, Chapter 6,
pp 149−171.
(8) Murray, S.; Briasoulis, E.; Linardou, H.; Bafaloukos, D.;
Papadimitriou, C. Cancer Treat. Rev. 2012, 38, 890−903.
(9) Orr, G. A.; Verdier-Pinard, P.; McDaid, H.; Horwitz, S. B.
Oncogene 2003, 22, 7280−7295.
(10) Ferlini, C.; Gallo, D.; Scambia, G. Expert Opin. Invest. Drugs
2008, 17, 335−347.
(11) Matesanz, R.; Trigili, C.; Rodriguez-Salarichs, J.; Zanardi, I.;
Pera, B.; Nogales, A.; Fang, W. S.; Jimenez-Barbero, J.; Canales, A.;
Barasoain, I.; Ojima, I.; Diaz, J. F. Bioorg. Med. Chem. 2014, 22, 5078−
5090.
(12) Mita, A. C.; Denis, L. J.; Rowinsky, E. K.; DeBono, J. S.; Goetz,
A. D.; Ochoa, L.; Forouzesh, B.; Beeram, M.; Patnaik, A.; Molpus, K.;
Semiond, D.; Besenval, M.; Tolcher, A. W. Clin. Cancer Res. 2009, 15,
723−730.
(13) U.S. FDA Center for Drug Evaluation and Research. Approval
Package for Jevtana. June 17, 2010. Available from https://www.
accessdata.fda.gov/drugsatfda_docs/nda/2010/201023s000Approv.
pdf.
(14) Yang, C.-G.; Barasoain, I.; Li, X.; Matesanz, R.; Liu, R.-H.;
Sharom, F. J.; Yin, D. L.; Diaz, J. F.; Fang, W.-S. ChemMedChem 2007,
2, 691−701.
(33) Walker, R. C.; Crowley, M. F.; Case, D. A. J. Comput. Chem.
2008, 29, 1019−1031.
(34) Anandakrishnan, R.; Aguilar, B.; Onufriev, A. V. Nucleic Acids
Res. 2012, 40, W537−541.
(35) Maier, J. A.; Martinez, C.; Kasavajhala, K.; Wickstrom, L.;
Hauser, K. E.; Simmerling, C. J. Chem. Theory Comput. 2015, 11,
3696−3713.
(36) Kaus, J. W.; Pierce, L. T.; Walker, R. C.; McCammon, J. A. J.
Chem. Theory Comput. 2013, 9, 4131−4139.
(37) Cerutti, D. S.; Duke, R. E.; Darden, T. A.; Lybrand, T. P. J.
Chem. Theory Comput. 2009, 5, 2322−2338.
(38) Ryckaert, J.-P.; Ciccotti, G.; Berendsen, H. J. C. J. Comput. Phys.
1977, 23, 327−341.
(39) Case, D. A.; Berryman, J. T.; Betz, R. M.; Cerutti, D. S.;
Cheatham, T. E., III; Darden, T. A.; Duke, R. E.; Giese, T. J.; Gohlke,
H.; Goetz, A. W.; Homeyer, N.; Izadi, S.; Janowski, P.; Kaus, J.;
Kovalenko, A.; Lee, T. S.; LeGrand, S.; Li, P.; Luchko, T.; Luo, R.;
Madej, B.; Merz, K. M.; Monard, G.; Needham, P.; Nguyen, H.;
Nguyen, H. T.; Omelyan, I.; Onufriev, A.; Roe, D. R.; Roitberg, A.;
Salomon-Ferrer, R.; Simmerling, C. L.; Smith, W.; Swails, J.; Walker,
R. C.; Wang, J.; Wolf, R. M.; Wu, X.; York, D. M.; Kollman, P. A.
AMBER 2015, San Francisco, 2015.
(40) Andreu, J. M. In Microtubule Protocols, Methods in Molecular
Medicine; Zhou, J., Ed.; Humana Press: Totowa, NJ, 2007; Vol. 137,
Chapter 2, pp 17−28.
(15) Coderch, C.; Tang, Y.; Klett, J.; Zhang, S. E.; Ma, Y. T.; Wang,
S. R.; Matesanz, R.; Pera, B.; Canales, A.; Jimenez-Barbero, J.;
Morreale, A.; Diaz, J. F.; Fang, W. S.; Gago, F. Org. Biomol. Chem.
2013, 11, 3046−3056.
(16) Zhou, Q.; Li, Y.; Jin, J.; Lang, L.-W.; Zhu, Z.-X.; Fang, W.-S.;
Chen, X.-G. Biol. Pharm. Bull. 2012, 35, 2170−2179.
́
(41) Dıaz, J. F.; Barasoain, I.; Andreu, J. M. J. Biol. Chem. 2003, 278,
8407−8419.
́
(42) Buey, R. M.; Dıaz, J. F.; Andreu, J. M.; O’Brate, A.;
Giannakakou, P.; Nicolaou, K. C.; Sasmal, P. K.; Ritzen
Namoto, K. Chem. Biol. 2004, 11, 225−236.
́
, A.;
I
DOI: 10.1021/acs.jnatprod.7b00857
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
(43) Kowalski, R. J.; Giannakakou, P.; Gunasekera, S. P.; Longley, R.
E.; Day, B. W.; Hamel, E. Mol. Pharmacol. 1997, 52, 613−622.
(44) Joe, P. A.; Banerjee, A.; Luduena, R. F. Cell Motil. Cytoskeleton
̃
2008, 65, 476−486.
(45) Lowe, J.; Li, H.; Downing, K. H.; Nogales, E. J. Mol. Biol. 2001,
313, 1045−1057.
(46) Wang, H. B.; Zhang, J. Q.; Lv, G. Y.; Ma, J. B.; Ma, P. K.; Du, G.
Y.; Wang, Z. L.; Tian, J. W.; Fang, W. S.; Fu, F. H. PLoS One 2014, 9,
e114688.
(47) Wang, H. B.; Li, H. Y.; Zuo, M. X.; Zhang, Y.; Liu, H.; Fang, W.
S.; Chen, X. G. Cancer Lett. 2008, 268, 89−97.
J
DOI: 10.1021/acs.jnatprod.7b00857
J. Nat. Prod. XXXX, XXX, XXX−XXX