Syntheses and characterization of non-bisphosphonate quinoline derivatives as new FPPS inhibitors

Article abstract of DOI:10.1016/j.bbagen.2013.11.006

Background Farnesyl pyrophosphate synthase (FPPS) is a key regulatory enzyme in the biosynthesis of cholesterol and in the post-translational modification of signaling proteins. It has been reported that non-bisphosphonate FPPS inhibitors targeting its allosteric binding pocket are potentially important for the development of promising anti-cancer drugs. Methods The following methods were used: organic syntheses of non-bisphosphonate quinoline derivatives, enzyme inhibition studies, fluorescence titration assays, synergistic effect studies of quinoline derivatives with zoledronate, ITC studies for the binding of FPPS with quinoline derivatives, NMR-based HAP binding assays, molecular modeling studies, fluorescence imaging assay and MTT assays. Results We report our syntheses of a series of quinoline derivatives as new FPPS inhibitors possibly targeting the allosteric site of the enzyme. Compound 6b showed potent inhibition to FPPS without significant hydroxyapatite binding affinity. The compound showed synergistic inhibitory effect with active-site inhibitor zoledronate. ITC experiment confirmed the good binding effect of compound 6b to FPPS, and further indicated the binding ratio of 1:1. Molecular modeling studies showed that 6b could possibly bind to the allosteric binding pocket of the enzyme. The fluorescence microscopy indicated that these compounds could get into cancer cells. Conclusions Our results showed that quinoline derivative 6b could become a new lead compound for further optimization for cancer treatment. General significance The traditional FPPS active-site inhibitors bisphosphonates show poor membrane permeability to tumor cells, due to their strong polarity. The development of new non-bisphosphonate FPPS inhibitors with good cell membrane permeability is potentially important.

Full text of DOI:10.1016/j.bbagen.2013.11.006

Biochimica et Biophysica Acta 1840 (2014) 1051–1062
Contents lists available at ScienceDirect
Biochimica et Biophysica Acta
journal homepage: www.elsevier.com/locate/bbagen
Syntheses and characterization of non-bisphosphonate quinoline
derivatives as new FPPS inhibitors
Jinggong Liu a, Weilin Liu a, Hu Ge a, Jinbo Gao b, Qingqing He a, Lijuan Su a, Jun Xu a, Lian-quan Gu a,
Zhi-shu Huang a, Ding Li a,
⁎
a
School of Pharmaceutical Sciences, Sun Yat-sen University, Guangzhou University City, 132 Waihuan East Road, Guangzhou 510006, P. R. China
Department of Biology and Chemistry, City University of Hong Kong, 83 Tat Chee Avenue, Kowloon, Hong Kong, P. R. China
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 9 August 2013
Received in revised form 11 October 2013
Accepted 8 November 2013
Available online 15 November 2013
Background: Farnesyl pyrophosphate synthase (FPPS) is a key regulatory enzyme in the biosynthesis of cholester-
ol and in the post-translational modification of signaling proteins. It has been reported that non-bisphosphonate
FPPS inhibitors targeting its allosteric binding pocket are potentially important for the development of promising
anti-cancer drugs.
Methods: The following methods were used: organic syntheses of non-bisphosphonate quinoline derivatives,
enzyme inhibition studies, fluorescence titration assays, synergistic effect studies of quinoline derivatives with
zoledronate, ITC studies for the binding of FPPS with quinoline derivatives, NMR-based HAP binding assays,
molecular modeling studies, fluorescence imaging assay and MTT assays.
Results: We report our syntheses of a series of quinoline derivatives as new FPPS inhibitors possibly targeting the
allosteric site of the enzyme. Compound 6b showed potent inhibition to FPPS without significant hydroxyapatite
binding affinity. The compound showed synergistic inhibitory effect with active-site inhibitor zoledronate. ITC
experiment confirmed the good binding effect of compound 6b to FPPS, and further indicated the binding ratio
of 1:1. Molecular modeling studies showed that 6b could possibly bind to the allosteric binding pocket of the en-
zyme. The fluorescence microscopy indicated that these compounds could get into cancer cells.
Conclusions: Our results showed that quinoline derivative 6b could become a new lead compound for further op-
timization for cancer treatment.
Keywords:
Quinoline derivative
Bisphosphonate
Farnesyl pyrophosphate synthase
Allosteric site
Cancer
General significance: The traditional FPPS active-site inhibitors bisphosphonates show poor membrane perme-
ability to tumor cells, due to their strong polarity. The development of new non-bisphosphonate FPPS inhibitors
with good cell membrane permeability is potentially important.
© 2013 Elsevier B.V. All rights reserved.
1. Introduction
drugs show poor cell-membrane permeability to tumor cells and
other non-endocytic cells [5,14]. Several approaches have been taken
Bisphosphonates are an important class of drug molecules that are
widely used to treat the bone loss associated with osteoporosis [1],
Paget's disease [2], hypercalcemia [3], and metastatic bone disease [4].
The main bisphosphonates in clinical use are thought to act primarily
by inhibiting the farnesyl diphosphate synthase (FPPS), an important
enzyme in the biosynthesis of cholesterol and in the post-translational
modification of signaling proteins [5–10]. The inhibition results in
decreased prenylation of proteins such as Ras, Rho, Rac, and Rap, and
impairment of cell survival signaling pathway [11–13]. However, due
to the strong polarity and high hydrophilicity, the bisphosphonate
to optimize the pharmacokinetic properties of bisphosphonates, such
as reducing bone affinity with deoxy-bisphosphonates, or reducing
the polarity of the compounds by increasing side chain lipophilicity
[15–19]. In our previous work, we have developed a new type of
nitrogen-containing bisphosphonate analogs with a geranyl group
attached to one of phosphate groups showing nanomole inhibition
effect to FPPS with our newly developed assay method [20,21].
Recently, Jahnke group has reported a new type of non-
bisphosphonate inhibitors identified by fragment-based discovery
targeting a new allosteric pocket of FPPS [22], and quinolines and
salicylic derivatives have been shown as inhibitors of FPPS. Here, we
report our further design and synthesis of another series of quinoline
derivatives as new FPPS inhibitors. The combination of quinoline deriv-
ative 6b with zoledronate, an inhibitor targeting active site of FPFS,
showed higher inhibition effect than that of one individual inhibitor.
Molecular modeling study showed that 6b can bind well with the
Abbreviations: BPs, bisphosphonates; BSA, bovine serum albumin; DTT, DL-dithiothrei-
tol; FPPS, farnesyl pyrophosphate synthase; HAP, hydroxyapatite; MTT, 3-(4,5-dimethyl-
thiazol-2-yl)-2,5-diphenyl-tetrazolium bromide; Tris, tris(hydroxymethyl)aminomethane
⁎
Corresponding author. Tel./fax: +86 20 3994 3058.
E-mail address: liding@mail.sysu.edu.cn (D. Li).
0304-4165/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.bbagen.2013.11.006

1052
J. Liu et al. / Biochimica et Biophysica Acta 1840 (2014) 1051–1062
allosteric site of the FPPS. The fluorescence microscopy indicated that
these quinoline derivatives could get into cancer cells.
under reflux at 80 °C for 5 h. After evaporation of the solvent, the
crude product was purified with gel chromatography to give compound
4a as a white solid (2.62 g, 80%). 1H NMR (400 MHz, CDCl3) δ 10.57
(s, 1H), 8.14 (d, J = 8.0 Hz, 1H), 8.09 (dd, J = 8.0, 1.6 Hz, 1H), 8.05
(d, J = 8.4 Hz, 1H), 7.76 − 7.69 (m, 2H), 7.63 (s, 1H), 7.61 − 7.52
(m, 2H), 7.05 − 7.01 (m, 1H), 4.76 (s, 3H), 3.96 (s, 3H); 13C NMR
(100 MHz, CDCl3) δ 168.9, 157.2, 148.7, 146.3, 144.3, 134.2, 131.9,
130.1, 129.7, 126.5, 121.0, 120.8, 120.7, 117.6, 115., 104.3, 52.3, 47.7;
ESI-HRMS m/z: calcd for C18H15N2O2Cl [M-H]− 325.0749, found
325.0747.
48 mg LiOH (dissolved in 0.1 mL H2O) was added to a solution of
4a (1 mmol, 0.326 g) in THF/CH3OH (2.5 mL, 2.5 mL), and the mixture
was heated under reflux for 3 h. After cooling and neutralization
with 1 N HCl, the solution was filtered and dried to give a white
solid 4b (296 mg, 94.8%). 1H NMR (400 MHz, DMSO) δ 12.80
(s, 1H), 9.47–9.45 (m, 1H), 9.17 (s, 1H), 8.12 (d, J = 7.2 Hz, 3H),
7.90 − 7.88 (m, 2H), 7.86 − 7.80 (m, 2H), 7.42 − 7.38 (m, 1H),
6.83 (d, J = 8.4 Hz, 1H), 6.63 (t, J = 7.4 Hz, 1H), 4.84 (s, 3H); 13C
NMR (100 MHz, DMSO) δ 169.9, 157.3, 150.1, 145.8, 141.6, 134.5,
133.4, 131.8, 130.2, 129.0, 128.7, 128.1, 127.7, 124.6, 122.3, 120.5,
120.3, 114.7, 111.8, 110.8, 47.6; ESI-HRMS m/z: calcd for C17H13ClN2O2
[M-H]− 312.0756, found 312.0759.
2. Materials and methods
2.1. Materials
All the reagents used in the biological assay were analytical pure.
Chemical reagents used in the synthesis were of research grade or better
and were obtained from commercial sources. Zoledronate was
purchased from the Dalian Meilun Biotech Co., Ltd.
2.2. Organic syntheses of quinoline derivatives
1H and 13C NMR spectra were recorded using TMS as the internal
standard in DMSO-d6, CD3COCD3 or CDCl3 with a Bruker BioSpin
GmbH spectrometer at 400 MHz; Mass spectra (MS) were recorded
on a Shimadzu LCMS-2010A instrument with an ESI-ACPI mass selective
detector, and high resolution mass spectra (HRMS) were recorded on
Shimadzu LCMS-IT-TOF. Flash column chromatography was performed
with silica gel (200–300 mesh) purchased from Qingdao Haiyang
Chemical Co. Ltd. Organic syntheses of bisphosphonate derivatives
were shown in Schemes 1–4.
2.2.3. Synthesis of ethyl 2-((4-chloroquinolin-2-yl)methoxy)benzoate (5a)
and 2-((4-chloroquinolin-2-yl)methoxy)benzoic acid (5b)
2.2.1. General procedures for the preparation of intermediates 1, 2, and 3
The syntheses of intermediates were as those reported before
[23,24] with some modifications. Aniline (100 mmol, 9.31 g) or
1-naphthylamine (100 mmol, 14.3 g) and ethyl acetoacetate
(100 mmol, 13.0 g) or ethyl 4-chloroacetoacetate (100 mmol, 16.4 g)
in 100 mL PPA were heated at 130 °C for 1 h. After cooling and neutral-
ization, POCl3 was added to the solid product, and heated under reflux
at 100 °C for 3 h. The mixtures were cooled and poured into ice-water
mixture. After neutralization, it was filtered and dried to give the
intermediates 1, 2 and 3, which were used directly for the following
reaction without further purification.
Intermediate 1 (10 mmol, 2.11 g) was dissolved in 50 mL CH3CN.
After the addition of ethyl 2-hydroxybenzoate (11 mmol, 1.83 g),
followed by K2CO3 (10 mmol, 1.38 g), the mixture was heated at
90 °C for 5 h. The solution was poured into water, and filtered to give
a white solid 5a (2.32 g, 68%). 1H NMR (400 MHz, CDCl3) δ 8.25
(d, J = 9.2 Hz, 1H), 8.10 (d, J = 10.0 Hz, 2H), 7.88 (dd, J = 7.8,
1.8 Hz, 1H), 7.82 − 7.78 (m, 1H), 7.67 − 7.63 (m, 1H), 7.46
(m, 1H), 7.05 (dd, J = 16.0, 8.0 Hz, 2H), 5.44 (s, 2H), 4.46
(q, J = 7.2 Hz, 2H), 1.44 (t, J = 7.2 Hz, 3H); 13C NMR (100 MHz,
CDCl3) δ 181.1, 166.3, 157.8 157.5, 148.2, 143.7, 133.5, 132.0,
130.6, 129.2, 127.4, 125.9, 124.2, 121.11, 121.05, 119.4, 113.5,
71.2, 61.1, 14.4; ESI-HRMS m/z: calcd for C19H16NO3Cl [M + Na]+
364.0711, found 364.0703.
2.2.2. Synthesis of methyl 2-((2-(chloromethyl)quinolin-4-yl)amino)
benzoate (4a) and 2-((2-(chloromethyl)quinolin-4-yl)amino)benzoic
acid (4b)
The synthesis of 4a was similar to that reported previously [25].
Intermediate 1 (10 mmol, 2.11 g) was added to a solution of methyl
2-aminobenzoate (10 mmol, 1.51 g) in ethanol, followed by two
drops of concentrated hydrochloric acid. The mixture was heated
The hydrolysis procedure for 5a was the same as that for 4a, and the
desired product 5b was obtained as a white solid. 1H NMR (400 MHz,
DMSO) δ 12.92 (s, 1H), 8.23 (d, J = 8.4 Hz, 1H), 8.12–8.09 (m, 2H),
7.91 (t, J = 7.8 Hz, 1H), 7.80 − 7.74 (m, 2H), 7.56 − 7.51 (m, 1H),
7.29 (d, J = 8.0 Hz, 1H), 7.07 (t, J = 7.4 Hz, 1H), 5.47 (s, 3H);
Scheme 1. Syntheses of the intermediates 1, 2 and 3. Reagents and conditions: (a) PPA, 130 °C, 1 h; (b) POCl3, 100 °C, 3 h.

J. Liu et al. / Biochimica et Biophysica Acta 1840 (2014) 1051–1062
1053
Scheme 2. Syntheses of 4a–6b. Reagents and conditions: (a) methyl 2-aminobenzoate, C2H5OH, conc. HCl, 5 h; (b) THF/CH3OH, LiOH, 3 h; (c) ethyl 2-hydroxybenzoate, K2CO3, CH3CN,
90 °C, 5 h.
13C NMR (100 MHz, DMSO) δ 167.1, 158.1, 156.8, 147.7, 142.1, 133.3,
131.1, 131.0, 129.0, 128.0, 124.8, 123.6, 121.6, 120.9, 119.4, 114.1,
70.5; ESI-HRMS m/z: calcd for C17H12NO3Cl [M-H]− 312.0433, found
312.0433.
2.2.5. Synthesis of ethyl 2-((4-chlorobenzo[h]quinolin-2-yl)methoxy)
benzoate (7a) and 2-((4-chlorobenzo[h]quinolin-2-yl)methoxy)benzoic
acid (7b)
Intermediate 2 was reacted with ethyl 2-hydroxybenzoate following
the procedure described in Section 2.2.3, and the desired products 7a
and 7b were obtained as white solids. 7a: 1H NMR (400 MHz, CDCl3)
δ 9.34 (d, J = 8.0 Hz, 1H), 8.13 (d, J = 11.2 Hz, 2H), 7.95 − 7.90
(m, 3H), 7.78 − 7.72 (m, 2H), 7.49 − 7.45 (m, 1H), 7.14 (d, J = 8.4 Hz,
1H), 7.05 (t, J = 7.2 Hz, 1H), 5.56 (s, 3H), 4.47 (q, J = 7.2 Hz, 2H), 1.45
(t, J = 7.2 Hz, 3H); 13C NMR (100 MHz, CDCl3) δ 166.5, 157.6, 156.3,
146.8, 143.4, 133.8, 133.5, 132.0, 130.9, 128.8, 128.6, 127.9, 127.4, 125.0,
123.6, 121.1, 121.0, 120.9, 120.0, 113.6, 71.3, 61.1, 14.5; ESI-HRMS m/z:
calcd for C23H18NO3Cl [M + H]+ 392.1048, found 392.1041.
2.2.4. Synthesis of ethyl 2-((4-((2-(methoxycarbonyl)phenyl)amino)
quinolin-2-yl)methoxy)benzoate (6a) and 2-((2-((2-carboxyphenoxy)
methyl)quinolin-4-yl)amino)benzoic acid (6b)
Compound 5a was reacted with methyl 2-aminobenzoate following
the procedure for the synthesis of 4a to give the desired product 6a, and
the hydrolysis procedure for 6a was the same as that for 4a. 6a: 1H NMR
(400 MHz, CDCl3) δ 10.65 (s, 1H), 8.16 (d, J = 8.0 Hz, 1H), 8.07–8.03
(m, 2H), 7.91 (s, 1H), 7.84 (dd, J = 7.8, 1.8 Hz, 1H), 7.73 (m, 1H), 7.62
(d, J = 8.0 Hz, 1H), 7.59 − 7.56 (m, 1H), 7.49 − 7.45 (m, 1H),
7.45 − 7.40 (m, 1H), 7.07 (d, J = 8.4 Hz, 1H), 6.99 (td, J = 8.6,
0.8 Hz, 2H), 5.43 (s, 2H), 4.24 (q, J = 7.2 Hz, 2H), 3.95 (s, 3H), 1.25
(t, J = 7.2 Hz, 3H); 13C NMR (100 MHz, CDCl3) δ 169.0, 166.0, 158.1,
157.6, 145.9, 144.3, 134.5, 133.4, 131.6, 131.6, 129.8, 129.2, 126.0,
121.1, 120.9, 120.8, 120.6, 120.3, 117.7, 115.4, 113.6, 103.1, 71.4,
60.7, 52.2, 14.2; ESI-HRMS m/z: calcd for C27H24N2O5 [M + H]+
457.1758, found 457.1729.
7b: 1H NMR (400 MHz, DMSO) δ 9.18 (dd, J =8.0, 4.0 Hz, 1H),
8.27 (s, 1H), 8.09 (s, 3H), 7.82 (dd, J = 6.2, 3.4 Hz, 2H), 7.64
(d, J = 6.4 Hz, 1H), 7.42 (t, J = 7.6 Hz, 1H), 7.24 (d, J = 8.0 Hz, 1H),
7.01 (t, J = 7.4 Hz, 1H), 5.53 (s, 2H); 13C NMR (100 MHz, DMSO)
δ 168.2, 157.2, 156.2, 145.9, 142.0, 133.3, 131.6, 130.4, 130.0, 129.0,
128.8 128.1, 127.6, 124.3, 122.6, 120.7, 120.2, 113.9, 70.6; ESI-HRMS
m/z: calcd for C21H14NOCl [M-H]− 362.0589, found 362.0603.
6b: 1H NMR (400 MHz, DMSO) δ 10.99 (s, 1H), 8.22 (d, J = 8.0 Hz,
1H), 8.05 (d, J = 8.0 Hz, 1H), 7.97 (d, J = 8.4 Hz, 1H), 7.83
(t, J = 7.4 Hz, 1H), 7.72 − 7.65 (m, 3H), 7.61 (d, J = 3.2 Hz, 2H),
7.50 − 7.46 (m, 2H), 7.27 (d, J = 8.4 Hz, 1H), 7.20 − 7.16 (m, 1H),
7.05 (t, J = 7.4 Hz, 1H), 5.49 (s, 3H); 13C NMR (100 MHz, DMSO)
δ 174.7, 172.3, 162.4, 161.8, 152.3, 151.5, 147.4, 139.2, 138.3, 137.1,
136.2, 135.9, 132.4, 1315, 127.6, 127.5, 126.4, 126.4, 124.8, 124.8,
124.3, 119.8, 106.0, 75.5; ESI-HRMS m/z: calcd for C24H18N2O5
[M + H]+ 415.1288, found 415.1269.
2.2.6. Synthesis of methyl 4-chloro-2-((4-((4-methoxyphenyl)amino)
benzo[h]quinolin-2-yl)methoxy)benzoate (8a) and 4-chloro-2-((4-((4-
methoxyphenyl)amino)benzo[h]quinolin-2-yl)methoxy)benzoic acid (8b)
Intermediate 2 was reacted with methyl 4-chloro-2-hydroxybenzoate
following the procedure described in Section 2.2.3, and the resulting
product methyl 4-chloro-2-((4-chlorobenzo[h]quinolin-2-yl)methoxy)
benzoate was then reacted with 4-methoxyaniline following the
procedure described in Section 2.2.2, to give the desired bright yellow
product 8a and yellow product 8b. 8a: 1H NMR (400 MHz, CDCl3) δ
Scheme 3. Syntheses of 13a and 13b. Reagents and conditions: (a) methyl 2-aminobenzoate, C2H5OH, conc. HCl, 5 h; (b) THF/CH3OH, LiOH, 3 h.

1054
J. Liu et al. / Biochimica et Biophysica Acta 1840 (2014) 1051–1062
Scheme 4. Syntheses of 7a–12b. Reagents and conditions: (a) ethyl 2-hydroxybenzoate, K2CO3, CH3CN, 90 °C, 5 h; (b) THF/CH3OH, LiOH, 3 h; (c) methyl 4-chloro-2-hydroxybenzoate,
K2CO3, CH3CN, 90 °C, 5 h; (d,e) 4-methoxyaniline, C2H5OH, conc. HCl, 5 h; (f) dimethyl malonate, NaH, THF, 0 °C, 10 h; (g) conc. HCl, 150 °C, 24 h; (h) ethyl 2-(diethoxyphosphoryl)acetate,
NaH, THF, 0 °C, 10 h; (i) triethyl phosphate; (j) dimethyl 5-aminoisophthalate, C2H5OH, conc. HCl, 5 h.
9.30 (d, J = 8.0 Hz, 1H), 7.86 (d, J = 7.2 Hz, 1H), 7.76 (q, J = 10.8 Hz,
3H), 7.72 − 7.63 (m, 2H), 7.29 (s, 1H), 7.23 − 7.19 (m, 3H), 6.95 (dd,
J = 13.0, 9.4 Hz, 3H), 6.67 (s, 1H), 3.83 (s, 3H), 3.68 (s, 3H); 13C NMR
(100 MHz, CDCl3) δ 174.8, 165.7, 158.7, 157.2, 139.4, 133.6, 132.8,
128.0, 127.6, 126.9, 125.3, 125.0, 120.8, 117.2, 115.2, 114.7, 114.5,
114.2, 101.0, 55.5, 51.8; ESI-HRMS m/z: calcd for C29H23ClN2O4
[M-H]− 498.1357, found 498.1346.
7.09 − 7.04 (m, 3H), 5.63 (s, 3H), 3.82 (s, 3H); ESI-HRMS m/z: calcd
for C28H21N2O4Cl [M-H]− 483.1117, found 483.1127.
2.2.7. Synthesis of dimethyl 2-((4-chlorobenzo[h]quinolin-2-yl)methyl)
malonate (9a) and 2-((4-chlorobenzo[h]quinolin-2-yl)methyl)malonic
acid (9b)
Sodium hydride (120 mg, 80%) was added to a solution of dimethyl
malonate (2 mmol, 0.264 g) dissolved in newly distilled THF at 0 °C.
After stirring for 30 min, a solution of intermediate 2 dissolved in
distilled THF was added dropwise to the mixture. After stirring for
10 h at 0 °C, the mixture was filtered, the solvent was removed, and
8b: 1H NMR (400 MHz, DMSO) δ 10.40 (s, 1H), 9.27 (s, 1H), 8.54
(d, J = 8.5 Hz, 1H), 8.17 (d, J = 4.8 Hz, 1H), 8.12 (d, J = 8.8 Hz, 1H),
7.91 − 7.87 (m, 2H), 7.77 (d, J = 8.4 Hz, 1H), 7.39 (d, J = 1.6 Hz,
1H), 7.31 (s, 1H), 7.29 (s, 1H), 7.17 (dd, J = 8.4, 1.2 Hz, 1H),

J. Liu et al. / Biochimica et Biophysica Acta 1840 (2014) 1051–1062
1055
the crude product was purified using silica gel chromatography to give
the desired brown product 9a with a yield of 62%. 1H NMR (400 MHz,
(s, 1H), 8.05 (s, 2H), 7.86 (d, J = 8.8 Hz, 1H), 7.79 (d, J = 6.0 Hz, 1H),
7.72 − 7.65 (m, 4H), 7.63 (s, 1H), 7.44 − 7.40 (m, 1H), 7.09
(d, J = 8.4 Hz, 1H), 6.98 (t, J = 7.4 Hz, 1H), 6.86 (s, 1H), 4.13
(q, J = 7.2 Hz, 2H), 3.90 (s, 6H), 1.22 (t, J = 7.2 Hz, 3H); 13C NMR
(100 MHz, CDCl3) δ 165.9, 158.0, 141.8, 133.4, 131.7, 130.9, 128.8,
128.1, 127.8, 126.9, 126.2, 125.5, 125.0, 120.4, 117.9, 116.4, 113.5,
103.8, 71.5, 60.7, 52.4, 14.2; ESI-HRMS m/z: calcd for C33H28N2O7
[M-H]− 563.1824, found 563.1842.
CDCl3)
δ 9.28 − 9.25 (m, 1H), 8.07 (d, J = 9.2 Hz, 1H), 7.90
(dd, J = 6.8, 2.4 Hz, 1H), 7.86 (d, J = 9.2 Hz, 1H), 7.75 − 7.69
(m, 2H), 7.53 (s, 1H), 4.54 (t, J = 7.6 Hz, 1H), 3.79 (s, 6H), 3.71
(d, J = 7.6 Hz, 2H); 13C NMR (100 MHz, CDCl3) δ 169.8, 156.4, 146.7,
142.6, 133.8, 131.1, 128.7, 128.2, 127.8, 127.3, 125.1, 122.9, 121.8,
120.8, 52.7, 49.8, 36.3; ESI-HRMS m/z: calcd for C19H16NO4Cl
[M + H]+ 358.0841, found 358.0861.
12b: 1H NMR (400 MHz, DMSO) δ 9.33 (d, J = 8.8 Hz, 1H), 8.55
(d, J = 9.2 Hz, 1H), 8.40 (s, 1H), 8.21 (m, 2H), 8.16 (s, 2H),
7.97 − 7.90 (m, 2H), 7.71 (dd, J = 7.6 Hz, 1.5 Hz, 1H), 7.57 − 7.50
(m, 1H), 7.43 (s, 1H), 7.25 (d, J = 8.3 Hz, 1H), 7.07 (t, J = 7.6 Hz,
1H); 13C NMR (100 MHz, DMSO) δ 166.6, 165.9, 156.3, 153.9, 152.2,
138.6, 134.2, 133.3, 132.9 131.3, 130.3, 129.2, 128.7, 128.0, 127.6,
124.0, 121.5, 121.4, 119.3, 114.8, 114.1, 101.2, 66.7; ESI-HRMS m/z:
calcd for C29H20N2O7 [M + H]+ 509.1343, found 509.1338.
2 mL concentrated hydrochloric acid was directly added to 9a
(1 mmol, 0.357 g) and heated at 150 °C for 24 h. The solution was neu-
tralized with 1 N NaOH. After filtration and drying, a brown product 9b
was obtained with nearly 100% yield. 9b: 1H NMR (400 MHz, Acetone) δ
9.35 − 9.32 (m, 1H), 8.11 (d, J = 9.2 Hz, 1H), 8.03 (t, J = 8.6 Hz, 2H),
7.77 (m, 3H), 4.43 (t, J = 7.2 Hz, 1H), 3.72 (d, J = 7.0 Hz, 3H); 13C
NMR (100 MHz, Acetone) δ 171.0, 159.0, 147.5, 142.8, 134.8, 131.9,
129.7, 129.3, 129.1 128.8, 128.3, 126.1 123.3, 123.1, 121.3, 50.3, 37.2;
ESI-HRMS m/z: calcd for C17H12NO4Cl [M-H]− 328.0382, found
328.0368.
2.2.11. Synthesis of methyl 2-(2-methylbenzo[h]quinolin-4-ylamino)
benzoate (13a) and 2-((2-methylbenzo[h]quinolin-4-yl)amino)benzoic
acid (13b)
2.2.8. Synthesis of ethyl 3-(4-chlorobenzo[h]quinolin-2-yl)-2-
(diethoxyphosphoryl)propanoate (10a) and 3-(4-chlorobenzo[h]
quinolin-2-yl)-2-phosphonopropanoic acid (10b)
Intermediate 3 was reacted with methyl 2-aminobenzoate following
the procedure described in Section 2.2.2, and the desired products 13a
and 13b were obtained as white solids. 13a: 1H NMR (400 MHz,
CDCl3) δ 10.36 (s, 1H), 9.34 (d, J = 7.2 Hz, 1H), 8.08 (dd, J = 8.0,
1.6 Hz, 1H), 7.99 (d, J = 9.2 Hz, 1H), 7.89 (d, J = 8.0 Hz, 1H), 7.78
(d, J = 9.2 Hz, 1H), 7.72 − 7.65 (m, 2H), 7.59 (d, J = 8.0 Hz, 1H),
7.48 (d, J = 7.2 Hz, 1H), 7.45 (s, 1H), 6.95 (t, J = 7.4 Hz, 1H), 3.96
(s, 3H), 2.77 (s, 3H); 13C NMR (100 MHz, CDCl3) δ 169.0, 158.0, 147.4,
145.4, 145.2, 134.1, 133.7 131.8, 131.6, 127.9, 127.6, 126.8, 126.2,
124.9, 119.7, 118.6, 117.3, 116.8, 114.8, 108.7, 52.2, 25.8; ESI-HRMS
m/z: calcd for C22H18N2O2 [M + H]+ 343.1441, found 343.1441.
13b: 1H NMR (400 MHz, DMSO) δ 13.63 (s, 1H), 10.85 (s, 1H), 9.38
(m, 1H), 8.49 (d, J = 9.2 Hz, 1H), 8.21 (dd, J = 9.2, 4.8 Hz, 2H), 8.11
(dd, J = 8.0, 1.2 Hz, 1H), 7.96 − 7.91 (m, 2H), 7.80 (td, J = 7.6,
1.2 Hz, 1H), 7.67 (d, J = 8.0 Hz, 1H), 7.56 (t, J = 7.2 Hz, 1H), 6.80
(s, 1H), 2.76 (s, 3H); 13C NMR (101 MHz, DMSO) δ 167.1, 154.0, 137.5,
137.1, 134.2, 133.8, 131.8, 130.2, 128.0, 127.9, 127.5, 127.5, 127.1,
126.7, 124.3, 122.9, 118.5, 102.8, 20.1; ESI-HRMS m/z: calcd for
Intermediate 2 was reacted with ethyl 2-(diethoxyphosphoryl)
acetate following the procedure described in Section 2.2.7. The desired
product 10a was obtained as a dark brown solid, and 10b was obtained
by hydrolyzing 10a with concentrated hydrochloric acid directly.
10a: 1H NMR (400 MHz, CDCl3) δ 9.26 (dd, J = 6.0, 2.8 Hz, 1H), 8.08
(d, J = 9.2 Hz, 1H), 7.92 − 7.90 (m, 1H), 7.85 (d, J = 9.2 Hz, 1H),
7.71 (m, 2H), 7.52 (s, 1H), 4.31 − 4.18 (m, 6H), 4.17 − 4.12 (m, 1H),
3.81 (m, 1H), 3.57 (m, 1H), 1.39 (dt, J = 10.8, 7.0 Hz, 6H), 1.23
(t, J = 7.2 Hz, 3H); 13C NMR (100 MHz, CDCl3) δ 169.0, 156.8, 146.7,
142.4, 133.7, 131.2, 128.6, 128.1, 127.7, 127.2, 125.2, 122.8, 121.8,
120.8, 62.9, 61.5, 44.1, 42.8, 34.5, 16.4, 14.1; ESI-HRMS m/z: calcd
for C22H25NO5PCl [M + H]+ 450.1232, found 450.1237.
10b: 1H NMR (400 MHz, DMSO) δ 9.22 (d, J = 4.8 Hz, 1H), 8.03
(s, 3H), 7.78 (s, 3H), 3.75 (d, J = 8.0 Hz, 1H), 3.69 − 3.63 (m, 2H);
ESI-HRMS m/z: calcd for C17H12NO4Cl [M-H]− 364.0147, found
364.0160.
C
21H16N2O2 [M + H]+ 329.1285, found 329.1273.
2.2.9. Synthesis of diethyl ((4-chlorobenzo[h]quinolin-2-yl)methyl)
phosphonate (11a) and ((4-chlorobenzo[h]quinolin-2-yl)methyl)
phosphonic acid (11b)
2.3. The enzyme inhibition studies
The enzyme inhibition studies were performed in flat bottom,
96-well plates using our newly developed non-radioactive methods.
100 ng of pure FPPS was incubated with or without inhibitor for
10 min in a final volume of 100 μL buffer, containing 50 mM Tris,
pH 7.5, 2 mM MgCl2, 1 mM DTT, 5 mg/mL BSA, and 100 μU/μL of
inorganic pyrophosphatase in each well. Then, the substrates were
added to start the reaction. Assays were terminated by the addition of
10 μL of 2.5% ammonium molybdate reagent (in 5 N H2SO4), 10 μL
of 0.5 M 2-mercaptoethanol and 5 μL of Eikonogen reagent (0.25 g of
sodium sulfite and 14.7 g of meta-bisulfite dissolved in 100 mL
water). The mixtures in plates were incubated with gentle mixing on
a plate shaker for 20 min. The absorbance was measured at 830 nm
using a Microplate Reader. IC50 was determined by a curve fitting of
relative activity versus inhibitor concentration using GraphPad Prism
5 Software.
Intermediate 2 was reacted with triethyl phosphate following the
procedure described in Section 2.2.7, and the desired products 11a
and 11b were obtained as brown solids. 11a: 1H NMR (400 MHz,
CDCl3) δ 9.27 (d, J = 8.0 Hz, 1H), 8.39 (s, 1H), 8.05 (s, 2H), 7.86
(d, J = 8.8 Hz, 1H), 7.79 (d, J = 6.0 Hz, 1H), 7.72 − 7.65 (m, 4H),
7.63 (s, 1H), 7.44 − 7.40 (m, 1H), 7.09 (d, J = 8.4 Hz, 1H), 6.98
(t, J = 7.4 Hz, 1H), 6.86 (s, 1H), 4.13 (q, J = 7.2 Hz, 2H), 3.90 (s, 6H),
1.22 (t, J = 7.2 Hz, 3H); ESI-HRMS m/z: calcd for C18H19NO3PCl
[M + H]+ 364.0864, found 364.0851.
11b: 1H NMR (400 MHz, DMSO) δ 9.21 (s, 1H), 8.08 (s, 3H), 7.79
(m, 3H), 3.55 (s, 1H), 3.49 (s, 1H); 13C NMR (100 MHz, DMSO) δ
154.8, 146.1, 140.8, 133.2, 130.2, 128.9, 128.4, 128.0, 127.4, 124.5,
122.8, 121.7, 120.2, 38.7; ESI-HRMS m/z: calcd for C14H11NO3PCl
[M-H]− 308.0238, found 308.0221.
2.2.10. Synthesis of dimethyl 5-((2-((2-(ethoxycarbonyl)phenoxy)methyl)
benzo[h]quinolin-4-yl)amino)isophthalate (12a) and 5-((2-((2-
carboxyphenoxy)methyl)benzo[h]quinolin-4-yl)amino)isophthalic acid
(12b)
Compound 7a was reacted with dimethyl 5-aminoisophthalate
following the procedure described in Section 2.2.2, and the desired
products 12a and 12b were obtained as yellow and brown solid, respec-
tively. 12a: 1H NMR (400 MHz, CDCl3) δ 9.27 (d, J = 8.0 Hz, 1H), 8.39
2.4. Fluorescence titration assay
Fluorescence titration assay was performed to evaluate the binding
affinity of these compounds to FPPS. The assay was carried out in
50 mM Tris buffer, pH 7.4, 5% glycerol, and 5 mM β-mercaptoethanol.
The excitation wavelength used in these experiments was 279 nm.
Emission spectra were scanned from 300 to 600 nm, with a 5 nm slit
width. For data analysis, the values measured for bound probe at the

1056
J. Liu et al. / Biochimica et Biophysica Acta 1840 (2014) 1051–1062
Table 1
The IC50 and Kd values for inhibition of FPPS by the compounds.
Compound
Zoledronate 4a
IC50 (μM) 0.2 4.10
4b
3.85
5a
5b
6a
6b
7a
3.51
7b
3.75
8a
7.86
8b
4.28
9a
9b
10a 10b
11a 11b
12a 12b
13a 13b
8.57
4.87 28.6
5.42 3.46
4.44 10.4
11.6 3.93 13.8
7.92 13.7
5.27 13.3
K
d (μM) 38.4
9
20.8
31.8 16.5 12.4
5.78 18.8
15.1
15.6
27.5
27.2 8.24 49.0 14.2
18.2 38.6 20.3 16.5
23.0 15.0
fluorescent emission peak of 538–540 nm were corrected for free
compounds/buffer and for any scattering occurred. These corrected
fluorescence enhancement data or reduced date were used to plot
against concentrations of compounds, and analyzed by nonlinear
regression to yield dissociation constants (Kd) and extrapolated
maximum fluorescence intensity (Fmax). Kd values were estimated by
fitting the titration data to the equation Y = [Fmax(X)] / [Kd + (X)].
and thermodynamic parameters of the reaction using Origin software
(OriginLab 7.0).
2.7. NMR-based HAP binding assay
HAP binding assay was performed on a 400 MHz Bruker DRX
spectrometer. Solutions (1.0 mL) of quinoline derivatives (50 mM in
aqueous buffer containing 50 mM Tris pH 7.5, and 10% D2O) were
divided into two aliquots. One aliquot was incubated with HAP
(2.0 mg) for 5 min, centrifuged, and measured by using NMR. The
other aliquot was treated identically, but not mixed with HAP.
2.5. The combination effect study of quinoline derivatives with zoledronate
The combination effect study of quinoline derivatives with
zoledronate was performed using the same enzyme inhibition assay.
The FPPS was incubated with zoledronate alone or with varying
concentrations of quinoline derivatives. The IC50 values for zoledronate
alone or zoledronate with different concentrations of quinoline
derivatives were calculated for the analyses of the combination effect
of these two types of compounds. The CompuSyn software [26,27]
was used to score the combination effect.
2.8. Molecular modeling study
To investigate the interacting mode between our synthetic
compounds and FPPS (PDB code: 3N6K), a molecular modeling study
utilizing the docking module named CDOCKER in Discovery Studio 2.5
(DS 2.5, Accelrys, Inc.) was carried out. The docked poses were clustered
and four ligand poses were selected as the starting pose for further
MD studies. MD simulation parameter files were prepared for each
ligand-receptor complex using LEaP in AMBER 12. The parameters for
ligands were adopted from the General Amber Force Field (GAFF),
while AMBERff12SB force field was applied for the receptors. Based on
the electrostatic potential (ESP) calculations at the ab initio HF/6-31G*
level, the partial atomic charges of ligands were calculated by using
the restricted electrostatic potential (RESP) fitting protocol implement-
ed in the Antechamber module of the AMBER 12 package. The
complexes were neutralized by adding sodium/chlorine counter ions,
and solvated in an octahedral box of TIP3P water molecules with solvent
layers 10 Å between the box edges and solute surface.
2.6. ITC study for the binding of quinoline derivatives with FPPS
ITC measurements were performed on a VP-ITC calorimeter. The
protein was dialyzed extensively against the buffer solution (50 mM
HEPES, pH 7.5, 25 mM NaCl, 2 mM DTT), and the ligand was dissolved
in the dialysis buffer. Calorimetric titrations were carried out at 25 °C.
A typical titration involved 25–30 injections of each ligand (4–10 μL
aliquot per injection) at 5-min intervals into the sample cell containing
the protein solution. The titration cell was stirred continuously at
307 r.p.m. The heat of the ligand dilution in the buffer alone was
subtracted from the titration data for each compound. The data were
analyzed to determine binding stoichiometry (N), binding affinity (Ka)
Fig. 1. The amino acid sequence of rat FPPS (accession code: EDM00670). Rat FPPS contains many tyrosines, tryptophans, and phenylalanines, such as Tyr10, Tyr58, Tyr204, Tyr349,
Phe206, Phe239, which let the enzyme have fluorescence.

J. Liu et al. / Biochimica et Biophysica Acta 1840 (2014) 1051–1062
1057
Fig. 4. (a) FPPS activity in the presence of zoledronate and compound 6b. (b) The IC50 of
zoledronate in the presence of varying concentration of 6b. FPPS was incubated with
zoledronate alone or with varying concentrations of compound 6b. The concentrations
of zoledronate were 50 nM, 100 nM, 150 nM, and 200 nM, and the concentrations of
compound 6b used were 0.4325 μM, 0.865 μM, 1.73 μM, and 3.46 μM.
Fig. 2. (a) The substrate binding sites and the allosteric binding site of FPPS. The left pocket
is the DMAPP or GPP binding site, the middle pocket is the IPP binding site, and the right
pocket is the allosteric binding site. (b) The computer model of allosteric pocket of rat
FPPS. Tyr10, Tyr56, Tyr204, and Tyr349 are located in this allosteric pocket.
solvent molecules with counter ions were allowed to move during a
2000-step minimization. All the atoms were allowed to be relaxed by
a 2000-step full minimization. After the minimization, each complex
was gradually heated from 0 to 300 K in 200 ps with solutes
Prior to the production step, the following protocol was applied,
using SANDER module in AMBER. The solutes were fixed, and the
Fig. 3. (a) Fluorescence titration of FPPS with zoledronate in the absence of 6b. (b) Fluorescence titration of FPPS with zoledronate in the presence of 6b. The FPPS used was 1 μM, and the
zoledronate used was 2 μM, 6 μM, 10 μM in the absence or presence of 2 μM 6b.

1058
J. Liu et al. / Biochimica et Biophysica Acta 1840 (2014) 1051–1062
1 atm and 300 K in the production step, using GPU accelerated PMEMD
in AMBER 12. The temperature was kept at 300 K by means of the weak-
coupling algorithm. The SHAKE algorithm was used to fix all covalent
bonds containing hydrogen atoms with a time step of 2 fs. The
particle-mesh Ewald (PME) method was performed to treat long-
range electrostatic interactions. A residue-based cutoff of 10 Å was
utilized for the non-covalent interactions.
The MM/PBSA method implemented in the AMBER 12 suite was
used to calculate the binding free energy between the receptor and
the ligand. The snapshots of last 10 ns from MD trajectories were col-
lected to calculate the binding free energies. The trajectory with a best
binding free energy was selected and its conformations were clustered
into 3 clusters. The representative conformation from the major cluster
was selected for binding mode analysis. The root mean square
deviations (RMSD) were calculated and plotted as time series for the
receptor backbone and the ligand, respectively.
2.9. MTT assay and fluorescence imaging assay
The types of tumor cells used in the present study included human
hepatocellular carcinoma cells (HepG2), human lung carcinoma cell
line (NCI), human cervical carcinoma cells (Hela), human leukemia
cells (A549 and HL60), and human normal epithelial cells (ECV-304).
The cells were seeded on 96-well plates (1.0 × 103/well), and exposed
to various concentrations of compounds. After 48 h of treatment at
37 °C in a humidified atmosphere of 5% CO2, the cells were centrifuged
and 10 mL of 5 mg/mL methyl thiazolyl tetrazolium (MTT) solution was
added to each well, and incubation was continued for another 4 h. The
cells in each well were then treated with dimethyl sulfoxide (DMSO)
(200 μL for each well), and the optical density (OD) was recorded at
570 nm. All compound doses were parallel tested in triplicate and the
highest dose was 100 μM, and the IC50 values were derived from the
mean OD values of the triplicate tests versus compound concentrations.
3. Results and discussion
Fig. 5. FPPS was titrated with 6b in ITC measurement. The data were analyzed to deter-
mine binding stoichiometry (N), binding affinity (Ka), and thermodynamic parameters
of the reaction using Origin software (OriginLab 7.0).
3.1. Organic syntheses of quinoline derivatives
The syntheses of the quinoline derivatives were as follow: the
starting materials aniline or 1-naphthylamine and ethyl acetoacetate
or ethyl 4-chloroacetoacetate were used to synthesize the intermedi-
ates as described previously [28]. These intermediates were reacted
with the aromatic esters or aliphatic esters to give the quinoline esters.
constrained at a weak harmonic constraint of 5 kcal mol−1 Å−2. Finally,
periodic boundary dynamics simulations of 30 ns were carried out with
an NPT (constant composition, pressure, and temperature) ensemble at
Fig. 6. In vitro HAP experiment for the binding of FPPS with 6b. FPPS was incubated with 6b in the presence or absence of 2.0 mg HAP for 5 min, centrifuged, and measured by using NMR.

J. Liu et al. / Biochimica et Biophysica Acta 1840 (2014) 1051–1062
1059
Fig. 7. Molecular modeling study for the interaction between FPPS and 6b. (a) The binding mode of FPPS with 6b. (b) The RMSD in the presence and absence of 6b.
These esters were hydrolyzed with LiOH to generate the acid products.
The structures of the lead compounds were validated by using 1H NMR,
13C NMR, and HRMS. The quinoline esters and quinoline acids were then
used for the enzyme inhibition studies.
the good binding interactions between the FPPS and our synthetic
molecules.
3.4. The studies for the combination of quinoline derivative and zoledronate
3.2. The enzyme inhibition studies
To investigate the combination effects of our synthetic compounds
with FPPS active-site inhibitor zoledronate, the FPPS was incubated
with zoledronate alone or with varying concentrations of compound
6b. Zoledronate alone gave a dose-dependent inhibition of FPPS with
IC50 value of ~200 nM, and the addition of compound 6b apparently
accelerated this inhibition (Fig. 4a), especially the combination of
3.46 μM 6b and 200 nM zoledronate had nearly 90% inhibition of
FPPS, indicating the possibility for the combined use of these two
inhibitors. The IC50 value of zoledronate decreased from 200 nM to
66 nM in the presence of 6b (Fig. 4b). Then, the CompuSyn software
was used to calculate the CI value (Supporting Information). The CI
values for the combinations of low concentrations of zoledronate
(50 nM, 100 nM, 150 nM) with various concentrations of 6b
(0.4325 μM, 0.865 μM, 1.73 μM, 3.46 μM) were approximately 1,
indicating the possibility of the additive effect or antagonism. The
combinations of 200 nM zoledronate and various concentrations of 6b
had the CI values of less than 1, indicating the possibility of the
synergism. However, the real effect for the combination use of these
two inhibitors may be more thoroughly evaluated through long term
The designed quinoline esters and acids were incubated with FPPS,
assayed with our non-radioactive method [20], and zoledronic acid
was used as a control. As shown in Table 1, all these synthetic
compounds showed good inhibitory activity against FPPS with their
IC50 values ranging from 3.46 μM to 28.6 μM. Generally, the acid
derivatives showed higher inhibitory effect than their ester derivatives.
It is possible that the hydrogen bonding interactions or ionic interac-
tions of the carboxylic acid made strong contribution to the binding.
The aromatic core could have hydrophobic interactions and/or π–π
stacking interactions with surrounding amino acids in the binding
pocket of FPPS. Compounds 6b and 7a displayed good inhibitory activity
with their IC50 values of 3.46 μM and 3.51 μM, respectively.
3.3. Fluorescence titration studies
In order to further confirm the binding between the FPPS and our
synthetic compounds, fluorescence titration experiment was per-
formed. Rat FPPS has some tyrosines and tryptophans with aromatic
side chains, which let the enzyme have fluorescence [29–31]. Based
on the amino acid sequence (Fig. 1) and the computer model of
three important binding pockets of rat FPPS (Fig. 2a and b), we found
that several residues including Tyr10, Tyr56, Tyr204, and Tyr349 are
located around the allosteric site. The binding of small molecule to the
enzyme may change hydrophilic and hydrophobic environment around
these amino acids, resulting in fluorescence absorbance changes. The
fluorescence titration experiment was carried out following that
described with minor modification [8,32]. The zoledronate was used
as a control. The dissociation constant Kd values (Table 1) were obtained
through data fitting using nonlinear regression analysis with equation
Y = [Fmax(X)] / [Kd + (X)]. Our result confirmed the binding interac-
tions between the FPPS and our synthetic molecules. The combination
of zoledronate and 6b decreased the fluorescence absorbance of FPPS
compared to that with zoledronate alone (Fig. 3). The Kd value for inhi-
bition of FPPS with zoledronate alone was 9.0 μM, and it was decreased
to 7.2 μM in the presence of 2 μM 6b. Similarly, the Kd value for 6b was
5.78 μM, and it was decreased to 4.84 μM in the presence of 2 μM
zoledronate. It should be noted that the calculated Kd values are related
with hydrophilic and hydrophobic environmental changes of the
surrounding aromatic residues, while IC50 values indicated inhibitory
effects. Therefore, they are not necessarily correlated, but both support
Table 2
The cytotoxic effects of the quinoline derivatives on cancer cells.
Compound
IC50 (μM)
HepG2
NCI
Hela
HL60
A549
ECV304
4a
4b
5a
5b
6a
6b
7a
7b
8a
8b
9a
9b
10a
10b
11a
11b
12a
12b
13a
13b
28.6
30.2
93.8
25.9
45.8
26.2
49.6
49.6
52.1
85.8
65.6
28.8
18.6
58.4
23.3
16.6
60.3
85.8
40.0
96.2
37.9
44.6
78.7
37.2
54.5
36.3
50.4
92.8
27.8
82.8
90.2
47.5
28.3
62.6
27.2
65.1
66.8
82.2
84.4
45.5
31.8
50.1
70.3
73.2
50.0
20.5
35.7
77.3
36.6
75.3
57.5
30.2
25.0
33.0
29.1
47.7
38.8
73.6
20.2
80.6
20.2
44.2
36.6
30.6
12.2
16.4
19.4
80.3
31.1
37.6
56.8
20.7
7.16
44.4
8.09
25.6
15.8
43.2
17.6
34.5
29.9
18.4
67.4
41.1
15.8
17.4
27.4
76.8
16.8
43.3
73.5
65.3
24.3
83.6
33.2
78.7
20.2
25.8
72.1
67.4
26.0
70.8
N100
60.6
99.7
70.7
N100
N100
N100
N100
N100
N100
65.3
N100
N100
N100
N100
N100
N100
N100

1060
J. Liu et al. / Biochimica et Biophysica Acta 1840 (2014) 1051–1062
Fig. 8. Uptake of 6a and 6b by human Hela cells. Cells were incubated with compound 6a or 6b (10 μM) for 6 h, and were subsequently examined via fluorescent microscopy. Control cells
under phase-contrast microscopy and under fluorescent microscopy are also shown. (a) Hela cells were treated with DMSO for 6 h; (b) Hela cells were treated with 6a for 6 h; (c) Hela
cells were treated with 6b for 6 h.
monitoring their injection in animal experiments in the future.
Although the IC50 value of compound 6b (3.46 μM) is ~17 times higher
than that of zoledronate, considering tumor cell permeability problem
of zoledronate, the combined use of these two types of compounds
might be more effective for cancer treatment.
resulting net enthalpy change was then determined by subtracting the
value of dilution control. From Fig. 5, we determined a binding stoichi-
ometry of approximately 1:1 for the binding of 6b to one subunit of
the FPPS homodimer. The binding of 6b to FPPS was found to be exo-
thermic, and the binding enthalpy and entropy in this entropy-driven
binding event were determined to be −1.64 kcal mol−1 and 18.8 cal
mol−1 K−1, respectively. The resulting Kd value was determined to be
4.8 μM, which is consistent with the measured IC50 value (3.46 μM)
with the enzymatic assay. The free energy (ΔG) was determined to
be −7.25 kcal mol−1, indicating that the binding of 6b to FPPS is
exothermic.
3.5. ITC study for the binding of quinoline derivatives with FPPS
In order to gain insight details on the binding affinity and the binding
ratio of quinoline derivatives for FPPS, ITC experiment was performed to
characterize the energy of molecular binding interactions [33–35]. The
titration of FPPS with compound 6b was carried out, with FPPS and 6b
both dissolved in the same dialysis buffer. The titration results including
corresponding enthalpy values are shown in Fig. 5, and small aliquots of
6b solution were added from the syringe into the microcalorimeter cells
containing FPPS. The ligand dilution control experiment was also
carried out, and the same 6b solution was added from the syringe into
the dialysis buffer without FPPS in the cell. We found that the enthalpy
change for ligand dilution was typically quite small in magnitude. The
3.6. NMR-based HAP binding assay
The traditional bisphosphonate compounds show strong binding to
hydroxyapatite (HAP) that is the major mineral component of bone,
making it less available for membrane penetration into tumor cells
[36]. In our experiment, the binding of quinoline derivatives to HAP
was determined by using 1H NMR. After the addition of 2.0 mg HAP,

J. Liu et al. / Biochimica et Biophysica Acta 1840 (2014) 1051–1062
1061
the 1H NMR signal (cyan line) was almost the same as that for the 6b
Appendix A. Supplementary data
alone (dark red line), as shown in Fig. 6. This result showed that the
compound 6b had almost no binding with bone material, which should
be more available for membrane penetration into tumor cells.
The CompuSyn software calculation of the CI value. This information
can be found in the online version at http://dx.doi.org/10.1016/
j.bbagen.2013.11.006.
3.7. Molecular modeling study
References
To investigate the interacting mode between our synthetic
compounds and FPPS (PDB code: 3N6K), a molecular modeling study
utilizing the docking program named AUTODOCK 4.0 package with
Discovery Studio 2.5 was carried out [37]. As shown in Fig. 7a,
compound 6b can well occupy the allosteric pocket of FPPS. The carbox-
ylate groups connected with two benzene rings show strong hydrogen-
bonding interactions with Tyr49, Gly47, Lys48 and Asn50, and Lys48
also has electrostatic interaction with the carboxylate group of 6b. The
oxygen atom located in the salicylic acid phenolic hydroxyl group also
has the similar hydrogen-bonding interaction with Lys48, improving
the binding affinity of 6b to this pocket. The quinoline core penetrates
deeply into the pocket, and the introduction of benzene ring contributes
π–π stacking interactions with Phe197 and Phe230, and hydrophobic
interactions with Tyr49, Gly47 and Lys48.
[1] R. Russell, P. Croucher, M. Rogers, Bisphosphonates: pharmacology, mechanisms of
action and clinical uses, Osteoporos. Int. 9 (1999) S66–S80.
[2] R.E. Coleman, The role of bisphosphonates in breast cancer, Breast 13 (Suppl. 1)
(2004) S19–S28.
[3] S. Sun, C.E. McKenna, Farnesyl pyrophosphate synthase modulators: a patent review
(2006–2010), Expert. Opin. Ther. Pat. 21 (2011) 1433–1451.
[4] R.G.G. Russell, Bisphosphonates: from bench to bedside, Ann. N. Y. Acad. Sci. 1068
(2006) 367–401.
[5] B.M. Wasko, A. Dudakovic, R.J. Hohl, Bisphosphonates induce autophagy by
depleting geranylgeranyl diphosphate, J. Pharmacol. Exp. Ther. 337 (2011) 540–546.
[6] J.R. Green, Bisphosphonates: preclinical review, Oncologist 9 (2004) 3–13.
[7] M.J. Rogers, S. Gordon, H. Benford, F. Coxon, S. Luckman, J. Monkkonen, J. Frith,
Cellular and molecular mechanisms of action of bisphosphonates, Cancer 88
(2000) 2961–2978.
[8] P. Dürrschmidt, J. Mansfeld, R. Ulbrich-Hofmann, Differentiation between
conformational and autoproteolytic stability of the neutral protease from Bacillus
stearothermophilus containing an engineered disulfide bond, Eur. J. Biochem. 268
(2001) 3612–3618.
The root mean square deviations (RMSD) were calculated and plot-
ted as time series for the receptor backbone and the ligand, respectively,
as shown in Fig. 7b. The backbone RMSD of FPPS reached 2 Å within a
short time and stayed relatively stable during the whole 30 ns simula-
tion, indicating a global equilibrium state. The ligand 6b RMSD was
stable for the first 10 ns and increased to 2 Å afterwards, then slightly
declined. The trend of ligand RMSD is in good agreement with that of
the receptor, which may imply a good ligand-induced fit of receptor.
[9] M.J. Fischer, S. Meyer, P. Claudel, M. Bergdoll, F. Karst, Identification of a lysine
residue important for the catalytic activity of yeast farnesyl diphosphate synthase,
Protein J. 30 (2011) 334–339.
[10] K.L. Kavanagh, K. Guo, J.E. Dunford, X. Wu, S. Knapp, F.H. Ebetino, M.J. Rogers, R.G.G.
Russell, U. Oppermann, The molecular mechanism of nitrogen-containing
bisphosphonates as antiosteoporosis drugs, Proc. Natl. Acad. Sci. 103 (2006)
7829–7834.
[11] R.G.G. Russell, Z. Xia, J.E. Dunford, U. Oppermann, A. Kwaasi, P.A. Hulley, K.L.
Kavanagh, J.T. Triffitt, M.W. Lundy, R.J. Phipps, Bisphosphonates, Ann. N. Y. Acad.
Sci. 1117 (2007) 209–257.
[12] C.M. Shipman, M.J. Rogers, J.F. Apperley, R.G.G. Russell, P.I. Croucher, Bisphosphonates
induce apoptosis in human myeloma cell lines: a novel anti-tumour activity,
Br. J. Haematol. 98 (1997) 665–672.
3.8. Cytotoxic effects on cancer cells and cell localization
[13] V. Stresing, F. Daubiné, I. Benzaid, H. Mönkkönen, P. Clézardin, Bisphosphonates in
cancer therapy, Cancer Lett. 257 (2007) 16–35.
MTT assays were performed to evaluate the cytotoxic effects of the
quinoline derivatives on cancer cells. Most compounds showed mild
cytotoxic effects to cancer cells, and the best inhibition activity was
found for compounds 10a and 11a with their IC50 values of 7.16 μM
and 8.09 μM for HL60 cells. The IC50 value of 6b ranged from 16.4 μM
to 36.3 μM for different cancer cells. These quinoline derivatives
showed nearly no cytotoxic effects to human normal vascular endothe-
lial cell ECV304, with their IC50 values mostly more than 100 μM
(Table 2).
Fluorescence imaging is a very important technique for biological
studies and clinical applications due to high temporal and spatial reso-
lutions. In the present study, fluorescent microscopy was performed
to assess the cellular uptake of the fluorescent compounds 6b and 6a.
Hela cells were incubated with compound 6b or 6a (10 μM) for 6 h,
and were subsequently examined via confocal fluorescent microscopy.
As shown in Fig. 8, the fluorescence microscopy clearly showed that
compounds 6b and 6a can penetrate Hela cells.
In summary, we designed and synthesized a series of non-
bisphosphonate quinoline derivatives, which were found to be new
FPPS inhibitors. Molecular modeling and ITC studies indicated that
compound 6b can well occupy the allosteric pocket of FPPS with a 1:1
binding stoichiometry. None of these compounds showed measurable
affinity to hydroxyapatite, indicating their low affinity to bone mineral
material. Compound 6b showed synergistic effect with zoledronate for
FPPS inhibition, indicating the possibility for the combined use of
these two inhibitors. Compound 6b could become a lead compound
for further development for the treatment of cancer.
[14] D. Glick, S. Barth, K.F. Macleod, Autophagy: cellular and molecular mechanisms,
J. Pathol. 221 (2010) 3–12.
[15] Y. Zhang, R. Cao, F. Yin, M.P. Hudock, R.-T. Guo, K. Krysiak, S. Mukherjee, Y.-G. Gao, H.
Robinson, Y. Song, J.H. No, K. Bergan, A. Leon, L. Cass, A. Goddard, T.-K. Chang, F.-Y.
Lin, E.V. Beek, S. Papapoulos, A.H.J. Wang, T. Kubo, M. Ochi, D. Mukkamala, E.
Oldfield, Lipophilic bisphosphonates as dual farnesyl/geranylgeranyl diphosphate
synthase inhibitors: an X-ray and NMR investigation, J. Am. Chem. Soc. 131
(2009) 5153–5162.
[16] P.G. Fournier, F. Daubine, M.W. Lundy, M.J. Rogers, F.H. Ebetino, P. Clezardin,
Lowering bone mineral affinity of bisphosphonates as a therapeutic strategy to op-
timize skeletal tumor growth inhibition in vivo, Cancer Res. 68 (2008) 8945–8953.
[17] J.M. Rondeau, F. Bitsch, E. Bourgier, M. Geiser, R. Hemmig, M. Kroemer, S. Lehmann,
P. Ramage, S. Rieffel, A. Strauss, J.R. Green, W. Jahnke, Structural basis for the excep-
tional in vivo efficacy of bisphosphonate drugs, ChemMedChem 1 (2006) 267–273.
[18] R. Cao, C.K. Chen, R.T. Guo, A.H. Wang, E. Oldfield, Structures of a potent phenylalkyl
bisphosphonate inhibitor bound to farnesyl and geranylgeranyl diphosphate
synthases, Proteins 73 (2008) 431–439.
[19] M.S. Marma, Z. Xia, C. Stewart, F. Coxon, J.E. Dunford, Synthesis and biological
evaluation of r -halogenated bisphosphonate and phosphonocarboxylate analogues
of risedronate, J. Med. Chem. 50 (2007) 9.
[20] J. Gao, X. Chu, Y. Qiu, L. Wu, Y. Qiao, J. Wu, D. Li, Discovery of potent inhibitor for
farnesyl pyrophosphate synthase in the mevalonate pathway, Chem. Commun.
(Camb.) 46 (2010) 5340–5342.
[21] J. Gao, J. Liu, Y. Qiu, X. Chu, Y. Qiao, D. Li, Multi-target-directed design, syntheses, and
characterization of fluorescent bisphosphonate derivatives as multifunctional
enzyme inhibitors in mevalonate pathway, Biochim. Biophys. Acta Gen. Subj. 1830
(2013) 3635–3642.
[22] W. Jahnke, J.-M. Rondeau, S. Cotesta, A. Marzinzik, X. Pellé, M. Geiser, A. Strauss, M.
Götte, F. Bitsch, R. Hemmig, C. Henry, S. Lehmann, J.F. Glickman, T.P. Roddy, S.J. Stout,
J.R. Green, Allosteric non-bisphosphonate FPPS inhibitors identified by fragment-
based discovery, Nat. Chem. Biol. 6 (2010) 660–666.
[23] L. Lengyel, T.Z. Nagy, G. Sipos, R. Jones, G. Dormán, L. Ürge, F. Darvas, Highly efficient
thermal cyclization reactions of alkylidene esters in continuous flow to give
aromatic/heteroaromatic derivatives, Tetrahedron Lett. 53 (2012) 738–743.
[24] R.M. Gengan, P. Pandian, C. Kumarsamy, P.S. Mohan, Convenient and Efficient
microwave-assisted synthesis of a methyl derivative of the fused indoloquinoline
alkaloid cryptosanguinolentine, Molecules 15 (2010) 3171–3178.
[25] A. Avetisyan, I. Aleksanyan, L. Ambartsumyan, Synthesis of 6, 8-substituted
4-(hydroxyphenylamino)-and 4-(aminophenylamino)-2-methylquinolines,
Russ. J. Org. Chem. 43 (2007) 1048–1051.
Acknowledgements
We thank Sun Yat-sen University and the National Natural Science
Foundation of China (Grant 21242010) for the financial support of
this study.
[26] T.C. Chou, Theoretical basis, experimental design, and computerized simulation of
synergism and antagonism in drug combination studies, Pharmacol. Rev. 58
(2006) 621–681.

1062
J. Liu et al. / Biochimica et Biophysica Acta 1840 (2014) 1051–1062
[27] T.C. Chou, Drug combination studies and their synergy quantification using the
Chou–Talalay method, Cancer Res. 70 (2010) 440–446.
[28] Z.-Q. Liu, S.-T. Zhuo, J.-H. Tan, T.-M. Ou, D. Li, L.-Q. Gu, Z.-S. Huang, Facile syntheses of
disubstituted bis(vinylquinolinium)benzene derivatives as G-quadruplex DNA
binders, Tetrahedron 69 (2013) 4922–4932.
[32] M.L. Colgrave, D.J. Craik, Thermal, chemical, and enzymatic stability of the cyclotide
kalata B1: the importance of the cyclic cystine knot†, Biochemistry 43 (2004)
5965–5975.
[33] M.L. Doyle, Characterization of binding interactions by isothermal titration
calorimetry, Curr. Opin. Biotechnol. 8 (1997) 31–35.
[29] F.H. Ebetino, A.M. Hogan, S. Sun, M.K. Tsoumpra, X. Duan, J.T. Triffitt, A.A. Kwaasi, J.E.
Dunford, B.L. Barnett, U. Oppermann, M.W. Lundy, A. Boyde, B.A. Kashemirov, C.E.
McKenna, R.G. Russell, The relationship between the chemistry and biological
activity of the bisphosphonates, Bone 49 (2011) 20–33.
[30] K. Ohno, K. Mori, M. Orita, M.T., Computational Insights into Binding of
Bisphosphates to Farnesyl Pyrophosphate Synthase Current Medicinal Chemistry,
18 (2011) 220–233.
[34] B. Pagano, C. Giancola, Energetics of quadruplex-drug recognition in anticancer
therapy, Curr. Cancer Drug Targets 7 (2007) 520–540.
[35] B. Pagano, C.A. Mattia, C. Giancola, Applications of isothermal titration calorimetry in
biophysical studies of G-quadruplexes, Int. J. Mol. Sci. 10 (2009) 2935–2957.
[36] W. Jahnke, C. Henry, An in vitro assay to measure targeted drug delivery to bone
mineral, ChemMedChem 5 (2010) 770–776.
[37] G.M. Morris, D.S. Goodsell, R.S. Halliday, R. Huey, W.E. Hart, R.K. Belew, A.J. Olson,
Automated docking using a Lamarckian genetic algorithm and an empirical binding
free energy function, J. Comput. Chem. 19 (1998) 1639–1662.
[31] X. Chu, W. Yu, L. Wu, X. Liu, N. Li, D. Li, Effect of a disulfide bond on mevalonate
kinase, Biochim. Biophys. Acta 1774 (2007) 1571–1581.