New fluorescence-based high-throughput screening assay for small molecule inhibitors of tyrosyl-DNA phosphodiesterase 2 (TDP2)

Article abstract of DOI:10.1016/j.ejps.2018.03.021

Tyrosyl-DNA phosphodiesterase 2 (TDP2) repairs topoisomerase II (TOP2) mediated DNA damages and causes resistance to TOP2-targeted cancer therapy. Inhibiting TDP2 could sensitize cancer cells toward TOP2 inhibitors. However, potent TDP2 inhibitors with favorable physicochemical properties are not yet reported. Therefore, there is a need to search for novel molecular scaffolds capable of inhibiting TDP2. We report herein a new simple, robust, homogenous mix-and-read fluorescence biochemical assay based using humanized zebrafish TDP2 (14M_zTDP2), which provides biochemical and molecular structure basis for TDP2 inhibitor discovery. The assay was validated by screening a preselected library of 1600 compounds (Z′ ≥ 0.72) in a 384-well format, and by running in parallel gel-based assays with fluorescent DNA substrates. This library was curated via virtual high throughput screening (vHTS) of 460,000 compounds from Chembridge Library, using the crystal structure of the novel surrogate protein 14M_zTDP2. From this primary screening, we selected the best 32 compounds (2% of the library) to further assess their TDP2 inhibition potential, leading to the IC₅₀ determination of 10 compounds. Based on the dose-response curve profile, pan-assay interference compounds (PAINS) structure identification, physicochemical properties and efficiency parameters, two hit compounds, 11a and 19a, were tested using a novel secondary fluorescence gel-based assay. Preliminary structure-activity relationship (SAR) studies identified guanidine derivative 12a as an improved hit with a 6.4-fold increase in potency over the original HTS hit 11a. This study highlights the importance of the development of combination approaches (biochemistry, crystallography and high throughput screening) for the discovery of TDP2 inhibitors.

Full text of DOI:10.1016/j.ejps.2018.03.021

EuropeanJournalofPharmaceuticalSciences118(2018)67–79
Contents lists available at ScienceDirect
European Journal of Pharmaceutical Sciences
journal homepage: www.elsevier.com/locate/ejps
New fluorescence-based high-throughput screening assay for small molecule
inhibitors of tyrosyl-DNA phosphodiesterase 2 (TDP2)
Carlos J.A. Ribeiroa, Jayakanth Kankanalaa, Ke Shib, Kayo Kurahashib, Evgeny Kiselevc,
Azhar Ravjic, Yves Pommierc, Hideki Aiharab, Zhengqiang Wanga,⁎
a
Center for Drug Design, Academic Health Center, University of Minnesota, Minneapolis, MN 55455, United States
Department of Biochemistry, Molecular Biology and Biophysics, University of Minnesota, Minneapolis, MN 55455, United States
Developmental Therapeutics Branch and Laboratory of Molecular Pharmacology, Center for Cancer Research, National Cancer Institute, National Institutes of Health,
b
c
Bethesda, MD 20892, United States
A R T I C L E I N F O
A B S T R A C T
Keywords:
Tyrosyl-DNA phosphodiesterase 2 (TDP2) repairs topoisomerase II (TOP2) mediated DNA damages and causes
resistance to TOP2-targeted cancer therapy. Inhibiting TDP2 could sensitize cancer cells toward TOP2 inhibitors.
However, potent TDP2 inhibitors with favorable physicochemical properties are not yet reported. Therefore,
there is a need to search for novel molecular scaffolds capable of inhibiting TDP2. We report herein a new
simple, robust, homogenous mix-and-read fluorescence biochemical assay based using humanized zebrafish
TDP2 (14M_zTDP2), which provides biochemical and molecular structure basis for TDP2 inhibitor discovery.
The assay was validated by screening a preselected library of 1600 compounds (Z′ ≥ 0.72) in a 384-well format,
and by running in parallel gel-based assays with fluorescent DNA substrates. This library was curated via virtual
high throughput screening (vHTS) of 460,000 compounds from Chembridge Library, using the crystal structure
of the novel surrogate protein 14M_zTDP2. From this primary screening, we selected the best 32 compounds (2%
of the library) to further assess their TDP2 inhibition potential, leading to the IC50 determination of 10 com-
pounds. Based on the dose-response curve profile, pan-assay interference compounds (PAINS) structure iden-
tification, physicochemical properties and efficiency parameters, two hit compounds, 11a and 19a, were tested
using a novel secondary fluorescence gel-based assay. Preliminary structure-activity relationship (SAR) studies
identified guanidine derivative 12a as an improved hit with a 6.4-fold increase in potency over the original HTS
hit 11a. This study highlights the importance of the development of combination approaches (biochemistry,
crystallography and high throughput screening) for the discovery of TDP2 inhibitors.
Tyrosyl-DNA phosphodiesterase 2 (TDP2)
Anti-cancer
Fluorescence assay
High-throughput screening (HTS)
Virtual screening
Guanidines
1. Introduction
DNA cleavage. These complexes are transient as TOP2 reseals cleaved
DNA at the end of its catalytic cycle (Pommier et al., 2016). However,
Tyrosyl-DNA phosphodiesterase 2 (TDP2), also known as TNF re-
ceptor associated factor (TRAF) and TNF receptor associated protein
(TTRAP) (Pype et al., 2000) and ETS1-associated protein 2 (EAPII) (Pei
et al., 2003), was the first human 5′-tyrosyl DNA phosphodiesterase
identified (Cortes Ledesma et al., 2009). TDP2 plays a major role in
DNA repair by specifically cleaving the 5′tyrosyl-DNA phosphodiester
bond of stalled topoisomerase II (TOP2) cleavage complexes [reviewed
in (Menon and Povirk, 2016; Pommier et al., 2014)].
TOP2 resolves topological problems with double-stranded DNA
during normal physiological processes, such as transcription and re-
plication. Mechanistically, TOP2 acts by generating TOP2-DNA clea-
vage complexes (TOP2cc) featuring a covalent phosphotyrosine linkage
between its active site tyrosine and the 5′ phosphate end at the site of
under certain conditions, such as exposure to TOP2 poisons, TOP2cc
become abortive (Fortune and Osheroff, 2000; Nitiss, 2009; Pommier,
2013). Repair of abortive TOP2cc by cellular DNA repair machinery,
particularly TDP2, can lead to cancer resistance to TOP2 poisons. By
inhibiting TDP2, cancer cells could be sensitized toward treatment with
TOP2 poisons, resulting in therapeutic efficacy with much lower doses
of TOP2 poisons (Marchand et al., 2016).
Several lines of evidence support the potential clinical benefits of
specifically targeting TDP2 in cancer therapy: (i) TDP2-deleted cells
show hypersensitivity to TOP2 poisons (Cortes Ledesma et al., 2009;
Zeng et al., 2011); (ii) TDP2 knockout vertebrate cells show normal cell
growth and knockout mice are viable without noticeable pathology,
suggesting that TDP2 inhibitors could be well tolerated (Gomez-
⁎
Corresponding author.
E-mail address: wangx472@umn.edu (Z. Wang).
https://doi.org/10.1016/j.ejps.2018.03.021
Received 19 February 2018; Received in revised form 16 March 2018; Accepted 20 March 2018
Availableonline21March2018
0928-0987/©2018ElsevierB.V.Allrightsreserved.

C.J.A. Ribeiro et al.
EuropeanJournalofPharmaceuticalSciences118(2018)67–79
Herreros et al., 2013); (iii) the oncogenic role of overexpressed TDP2
has been established in several non-small-cell lung carcinoma cells (Li
et al., 2011), and TDP2 up-regulation attained by p53 gain-of-function
mutation has been linked to TOP2 poison resistance in human lung
cancer (Do et al., 2012); and (iv) tumor cells frequently lack parallel
DNA repair pathways, which makes them more vulnerable to certain
cancer chemotherapy (principle of synthetic lethality) (Curtin, 2012),
and more specifically to TOP2 inhibitors (Hoa et al., 2016; Maede et al.,
2014; Pommier et al., 2014).
In addition, TDP2 is also implicated in picornavirus and hepatitis B
virus (HBV) infections during steps where cleavage of a 5′-phospho-
tyrosine covalent bond occurs. In particular, picornavirus uses TDP2,
also known as VPg-unlinkase, to release the viral genomic RNA from
VPg protein necessary during RNA replication (Maciejewski et al.,
2016; Virgen-Slane et al., 2012). HBV uses the DNA repair machinery of
the infected cells to convert viral genomic relaxed circular DNA (RC-
DNA) into covalently closed circular DNA (cccDNA) where TDP2 could
play a role in releasing the viral P-protein from the 5′end of RC-DNA
(Koniger et al., 2014).
Since the discovery of TDP2 in 2009 (Cortes Ledesma et al., 2009),
only a few scaffolds have been identified possessing TDP2 inhibitory
activity (Laev et al., 2016): compound 1 (Ro 08-2750, Fig. 1) (Thomson
et al., 2013), deazaflavins (e.g. 2) (Hornyak et al., 2016; Marchand
et al., 2016; Raoof et al., 2013), isoquinoline-1,3-diones (e.g. 3)
(Kankanala et al., 2016), the triple inhibitors TOP1/TDP1/TDP2 in-
denoisoquinolines (e.g. 4) (Beck et al., 2016; Wang et al., 2017),
compounds 5 (NSC375986), 6 (NSC114532) (Kossmann et al., 2016)
and 7 (NSC111041) (Kont et al., 2016). Deazaflavin derivatives are the
only described TDP2 inhibitors with activities in the nanomolar range.
However, their use as molecular probes in studying cellular functions
and their potential as drug candidates are severely hindered by the poor
cell permeability. (Hornyak et al., 2016). Therefore, new scaffolds with
good potencies and desirable physicochemical properties are highly
desired.
The first reported methods developed to biochemically measure
TDP2 activity employed gel-based assays (Cortes Ledesma et al., 2009;
Gao et al., 2012; Zeng et al., 2011). In addition to using radiolabeled-
substrates, these assays can be expensive and time-consuming, and thus
may not be suitable for high-throughput screening (HTS). Colorimetric
assays using T5NPP (Adhikari et al., 2011) or NPPP (Raoof et al., 2013;
Thomson et al., 2013) as substrate have been developed, though diffi-
culties in achieving enzyme inhibition above 75% were observed when
NPPP was employed, and high enzyme concentration (30–36 nM) was
required for both chromogenic assays. Since TDP2 prefers more phy-
siologically relevant 5′-phosphotyrosyl oligonucleotides substrates over
the small compound surrogates (Gao et al., 2012), a new colorimetric
assay using a DNA substrate was reported (Thomson et al., 2013).
However, this particular assay required the addition of calf intestinal
alkaline phosphatase (CIP) to cleave the phosphate group necessary for
the reaction development. Recently, Hornyak et al. reported a fluores-
cence-based assay using a 13-mer oligonucleotide substrate with a 5′-
tyrosine conjugated with FITC fluorophore and an enzyme concentra-
tion much lower than the chromogenic assays (50 pM) (Hornyak et al.,
2016). However, the TR-FRET nature of this assay required the addition
of trivalent metal ion sensor (Gyrasol technologies) to quench the
fluorescence of the substrate while stopping the reaction, resulting in
increased assay costs, and allowing only end-point quenched readings.
We report herein a new fluorescence-based assay allowing reading
in both continuous and quenched modality. With quenched reaction
protocol this new assay is amenable for HTS and requires low enzyme
concentration. In addition, the continuous reaction reading allows easy
detection of false positives due to the presence of fluorescent com-
pounds, as well as kinetic data collection (Acker and Auld, 2014). By
employing a humanized zebrafish protein (14M_zTDP2) developed by
our group, and whose crystal structure is included in this report, we
screened a library of 1600 compounds preselected via virtual high-
throughput screening (vHTS).
2. Materials and methods
2.1. Chemistry
All commercial chemicals were used as supplied unless otherwise
indicated. Flash chromatography was performed on
a Teledyne
Combiflash RF-200 with RediSep columns (silica) and indicated mobile
phase. All moisture sensitive reactions were performed under an inert
atmosphere of ultrapure argon with oven-dried glassware. 1H and 13C
NMR spectra were recorded on a Varian 600 MHz or Bruker 400 spec-
trometer. Mass data were acquired on an Agilent 6230 TOF LC/MS
spectrometer capable of ESI and APCI ion sources. All tested com-
pounds have a purity ≥95%.
2.1.1. General procedural for synthesis of 2, 2, 4-trimethyl
dihydroquinolines (10)
To a solution of corresponding aniline 8 (10 mmol) in acetone
(15 mL), was added catalytic InCl3 (5 mol%) and the resulting mixture
Fig. 1. Representative reported TDP2 inhibitors.
68

C.J.A. Ribeiro et al.
EuropeanJournalofPharmaceuticalSciences118(2018)67–79
was stirred at 50 °C for 12–24 h. The solvent was removed in vacuo and
the crude was dissolved in CH2Cl2, washed with Na2CO3 solution and
brine, dried over Na2SO4 and evaporated in vacuo to produce the crude
product. Purification of the crude product using Combi flash with
0–40% hexane in ethyl acetate as eluent furnished the desired product
in 40–65% yield. Adapted from (Li et al., 2015).
2.1.4. 2-((6-Methoxy-4-methylquinazolin-2-yl) amino) quinazolin-4(3H)-
one (12b) (Guiles et al., 2009)
A mixture of 11a (0.05 g, 0.22 mmol) and anthranilic acid (0.04 g,
0.26 mmol) in 1 mL of DMSO was heated at 170 °C until the starting
material disappeared. The reaction mixture was cooled, water was
added, and the resulting product was filtered and washed with excess
water. The crude mixture was recrystallized from chloroform to yield
the desired product as colorless crystals (0.046 g, 64%). 1H NMR
(600 MHz, DMSO‑d6) δ 13.60 (s, 1H), 11.11 (s, 1H), 8.04 (d, J = 7.7 Hz,
1H), 7.77–7.69 (m, 2H), 7.60 (dd, J = 9.1, 2.7 Hz, 1H), 7.47 (t,
J = 7.6 Hz, 2H), 7.30 (t, J = 7.4 Hz, 1H), 3.92 (s, 3H), 2.89 (s, 3H).
HRMS-ESI (+) m/z calculated for C18H16N5O2, 334.1304 [M + H]+;
found: 334.1307. Adapted from (Shikhaliev et al., 2003).
2.1.1.1. 6-Methoxy-2,2,4-trimethyl-1,2-dihydroquinoline
(10a). Yield:
40%; 1H NMR (600 MHz, DMSO‑d6) δ 6.59–6.52 (m, 2H), 6.41 (d,
J = 9.0 Hz, 1H), 5.38 (s, 1H), 5.31 (s, 1H), 3.64 (s, 3H), 1.88 (d,
J = 0.8 Hz, 3H), 1.16 (s, 6H).
2.1.1.2. 6-Chloro-2,2,4-trimethyl-1,2-dihydroquinoline (10c) (Li et al.,
2006). Yield: 65%; 1H NMR (600 MHz, CDCl3) δ 7.00 (s, 1H), 6.92
(d, J = 8.3 Hz, 1H), 6.36 (d, J = 7.7 Hz, 1H), 5.35 (s, 1H), 3.11 (s, 1H),
1.96 (s, 3H), 1.27 (s, 6H).
2.1.5. 1-(6-Methoxy-4-oxo-3,4-dihydroquinazolin-2-yl)guanidine (15)
To a solution of anthranilic acid (0.1 g, 0.59 mmol) in 1 mL of 10%
H2SO4, was added cyanoguanidine (0.075 g, 0.89 mmol) and refluxed
for 60 min. The hot solid was precipitated and washed with excess
water to furnish the desired product as colorless solid (0.077 g, 55%).
1H NMR (600 MHz, DMSO‑d6) δ 11.12 (s, 1H), 7.48 (s, 2H), 7.35 (d,
J = 2.7 Hz, 1H), 7.31 (d, J = 8.8 Hz, 1H), 7.22 (dd, J = 8.8, 2.7 Hz,
1H), 3.80 (s, 3H). HRMS-ESI (+) m/z calculated for C10H12N5O2,
234.0991 [M + H]+; found: 234.0994.
2.1.2. General procedural for synthesis of substituted quinazolines (11)
To a solution of 10 (2.46 mmol) in water (5 mL), was added cya-
noguanidine (2.46 mmol), concentrated HCl (1 mL) and the resulting
mixture was stirred under reflux for 4 h. The hot mixture was decanted
from oils, cooled and stirred continuously by treating with concentrated
KOH until a pH of 10–11 was obtained. The resulting precipitate was
filtered and washed several times with isopropanol followed by re-
crystallization from ethanol gave the desired product in 55–62% yield.
Adapted from (Rosowsky and Modest, 1972; Webb et al., 2003).
2.1.6. 4-Amino-5-(furan-2-yl)-2,4-dihydro-3H-1,2,4-triazole-3-thione
(18) (Koparır et al., 2004)
To a solution of KOH (0.67 g, 11.89 mmol) in methanol (10 mL),
was added furoic hydrazide (1.0 g, 7.92 mmol), CS2 (0.72 mL,
11.89 mmol) and stirred at room temperature for 24 h and then ether
was added and stirred the reaction mixture for another 2 h. The re-
sulting solid was filtered, washed with cold methanol and ether and air
dried. The solid was refluxed in excess of hydrazine hydrate (10 eq) for
5 h, cooled and poured into the acidic water (pH = 2). The solid was
filtered and recrystallized from ethanol to produce the desired com-
pound as colorless crystals (1.01 g, 70% over two steps). 1H NMR
(600 MHz, DMSO‑d6) δ 13.92 (s, 1H), 7.94 (dd, J = 1.8, 0.6 Hz, 1H),
7.39 (dd, J = 3.5, 0.6 Hz, 1H), 6.73 (dd, J = 3.5, 1.8 Hz, 1H), 5.83 (s,
2H). Adapted from (Khan et al., 2016).
2.1.2.1. 1-(6-Methoxy-4-methylquinazolin-2-yl) guanidine (11a). Yield:
55%; 1H NMR (600 MHz, DMSO‑d6) δ 7.51 (d, J = 9.0 Hz, 1H), 7.37 (d,
J = 9.1 Hz, 1H), 7.28 (s, 1H), 7.07 (s, 2H), 3.87 (s, 3H), 2.71 (s, 3H).
13C NMR (150 MHz, DMSO‑d6) δ 167.4, 159.3, 155.1, 145.9, 127.7,
125.2, 119.7, 104.3, 55.8, 22.0. HRMS-ESI (+) m/z calculated for
C
11H14N5O, 232.1198 [M + H]+; found: 232.1196.
2.1.2.2. 1-(6-Ethoxy-4-methylquinazolin-2-yl) guanidine (11b) (Brown,
1965). Yield: 62%; 1H NMR (600 MHz, DMSO‑d6)
7.50 (d,
δ
J = 9.1 Hz, 1H), 7.35 (dd, J = 9.1, 2.6 Hz, 1H), 7.27 (d, J = 2.6 Hz,
1H), 6.98 (s, 2H), 4.13 (q, J = 6.9 Hz, 2H), 2.69 (s, 3H), 1.38 (t,
J = 6.9 Hz, 3H). 13C NMR (100 MHz, DMSO‑d6) δ 167.0, 159.0, 154.0,
145.5, 127.3, 125.1, 119.4, 104.8, 63.6, 21.7, 14.7. HRMS-ESI (+) m/z
calculated for C12H16N5O, 246.1349 [M + H]+; found: 246.1340.
2.1.7. General procedure for the synthesis of 3-(furan-2-yl)-6-phenyl-
[1,2,4]triazolo[3,4-b][1,3,4]thiadiazole derivatives (19)
To a mixture of 18 (1.0 mmol) in POCl3 (5 mL), was added sub-
stituted benzoic acid (1.1 mmol) and refluxed for 6 h. The reaction
mixture was cooled and slowly poured into crushed ice with stirring
and neutralized with sodium bicarbonate. The precipitated solid was
filtered, washed with cold water, dried, and recrystallized from ethanol
to afford the desired compounds (19) as colorless crystals in 53–70%
yield. Adapted from (Khan et al., 2016).
2.1.2.3. 1-(6-Chloro-4-methylquinazolin-2-yl) guanidine hydrochloride
(11c) (Theiling and McKee, 1952). Yield: 58%; 1H NMR (600 MHz,
DMSO‑d6) δ 11.39 (s, 1H), 10.91 (s, 1H), 8.19 (s, 3H), 8.12 (d,
J = 9.0 Hz, 1H), 8.07 (d, J = 2.3 Hz, 1H), 7.71 (dd, J = 9.0, 2.3 Hz,
1H), 2.67 (s, 3H). 13C NMR (150 MHz, DMSO‑d6) δ 201.8, 155.6, 151.7,
136.8, 133.9, 131.2, 127.7, 126.9, 123.5, 29.2. HRMS-ESI (+) m/z
calculated for C10H11ClN5, 236.0697 [M + H]+; found: 236.0696.
2.1.7.1. 6-(3,4-Dimethylphenyl)-3-(furan-2-yl)-[1,2,4]triazolo[3,4-b]
[1,3,4] thiadiazole (19a). Yield: 53%; 1H NMR (600 MHz, DMSO‑d6) δ
8.04 (s, 1H), 7.84 (s, 1H), 7.78 (d, J = 7.6 Hz, 1H), 7.41 (t, J = 6.4 Hz,
2H), 6.82 (s, 1H), 2.35 (s, 3H), 2.33 (s, 3H). 13C NMR (150 MHz,
DMSO‑d6) δ 167.5, 145.4, 142.5, 140.1, 139.3, 138.1, 130.6, 127.8,
126.5, 124.8, 112.2, 111.8, 19.60, 19.18. HRMS-ESI (+) m/z calculated
for C15H13N4OS, 297.0810 [M + H]+; found: 297.0813.
2.1.3. 2-((6-Methoxy-4-methylquinazolin-2-yl) amino)-5, 6-dimethylpyrimidin-
4(3H)-one (12a)
To a solution of 11a (0.10 g, 0.43 mmol) in minimal hot DMSO, was
added ethyl 2-methylacetoacetate (0.07 mL, 0.52 mmol) and NaHCO3
(0.044 g, 0.52 mmol). The reaction mixture was heated at 170 °C until
the disappearance of starting material. The solution was cooled, and
water was added to precipitate the product which was collected by
filtration, washed with acetone and water and dried to afford 12a
(0.04 g, 30%) as a tan solid. 1H NMR (600 MHz, DMSO‑d6) δ 13.28 (s,
1H), 11.00 (s, 1H), 7.70 (d, J = 9.0 Hz, 1H), 7.60 (dd, J = 9.1, 2.7 Hz,
1H), 7.48 (d, J = 2.5 Hz, 1H), 3.93 (s, 3H), 2.87 (s, 3H), 2.18 (s, 3H),
1.90 (s, 3H). HRMS-ESI (+) m/z calculated for C16H18N5O2, 312.1460
[M + H]+; found: 312.1461. Adapted from (LaPorte et al., 2014).
2.1.7.2. 3-(Furan-2-yl)-6-(2-methoxyphenyl)-[1,2,4]triazolo[3,4-b]
[1,3,4]thiadiazole (19b) (Zhang et al., 2002). Yield: 70%; 1H NMR
(600 MHz, DMSO‑d6) δ 8.32 (dd, J = 7.9, 1.5 Hz, 1H), 8.02 (dd,
J = 1.6, 0.5 Hz, 1H), 7.74–7.56 (m, 1H), 7.41 (d, J = 3.5 Hz, 1H),
7.35 (d, J = 8.4 Hz, 1H), 7.21 (t, J = 7.6 Hz, 1H), 6.81 (dd, J = 3.4,
1.8 Hz, 1H), 4.06 (s, 3H). 13C NMR (150 MHz, DMSO‑d6) δ 162.5,
157.1, 154.8, 145.2, 140.2, 138.6, 134.4, 127.9, 121.7, 116.8, 113.1,
112.2, 111.3, 56.5. HRMS-ESI (+) m/z calculated for C14H11N4O2S,
299.0603 [M + H]+; found: 299.0606.
69

C.J.A. Ribeiro et al.
EuropeanJournalofPharmaceuticalSciences118(2018)67–79
2.2. Curation of screening compounds
frozen in liquid nitrogen for data collection. The crystals were screened,
and x-ray diffraction data collected at the Advanced Photon Source NE-
CAT 24 ID-E beamline at the wavelength of 0.979 Å. The diffraction
datasets were processed using HKL2000 (Otwinowski and Minor, 1997)
and XDS (Kabsch, 2010). Structure was solved by molecular replace-
ment using zebrafish TDP2 structure (PDB ID: 4F1H) as the search
model using PHASER program (McCoy et al., 2007). Three copies of
TDP2 monomers were located in the asymmetric unit. Initial refinement
of the molecular replacement solutions using REFMAC (Murshudov
et al., 1997) yield Rwork and Rfree of 37% and 41% respectively. Sub-
sequent refinement and model building were done using PHENIX
(Adams et al., 2010) and COOT (Emsley et al., 2010). Final Rwork and
Rfree are 18.3% and 20.9%, respectively. The summary of data collec-
tion and model refinement statistics is shown in Table S1 (Supple-
mentary material). Molecular graphic images were produced using
PYMOL (www.pymol.org). The atomic coordinates and structure factor
amplitudes have been deposited in the Protein Data Bank with acces-
sion code 6CA4.
Structure based virtual screening (VS) is routinely performed in
drug discovery involving in hit identification as a complementary ap-
proach to HTS which requires enormous cost, time and resources etc. A
computational docking tool “Glide” (Friesner et al., 2004; Schrödinger,
2014b) from Schrodinger suite 2014-3 was used to perform virtual
screening of larger commercial libraries which offers speed and accu-
racy in binding mode predictions and providing consistently high en-
richment at every level. The docking protocol involves several steps
which includes a) protein preparation: The crystal structure of
z2hTDP2cat (14M_zTDP2cat) was used for protein preparation in which
bond orders and missing side chains were corrected, polar hydrogens
are added, metal binding states were predicted, and the protein was
minimized (pdb code: 6CA4). b) ligand preparation involving applying
protonation and tautomeric states to the compounds from Chembridge
library using lig prep (Schrödinger, 2014a) and Epik; c) Grid genera-
tion: creating a grid around the metal ions (Mn2+) in the catalytic site
and imposing a pharmacophore constraint with the metal ion to im-
prove the chances of success in finding a hit molecule and d) docking of
small molecule library (460 K) to the prepared protein of interest. The
compound library (460 K) was first docked and ranked in Glide vHTS
mode. The top 115 K molecules (25%) from vHTS mode were then
docked and ranked using standard precision (SP) Glide (Friesner et al.,
2004). The resultant top 28,750 (25%) compounds from SP mode were
then docked and ranked using the more accurate and computationally
intensive extra-precision (XP) mode (Friesner et al., 2006). The top
7200 compounds were selected for visual inspection and by applying a
stringent constraint of physicochemical properties and synthetic tract-
ability, we created a small sub-set of 1600 compounds for purchasing
and biological evaluation. The screening protocol reported above was
routinely and effectively utilized in the structure based virtual
screening to aid hit-identification of various drug targets (Cain et al.,
2014; Ravindranathan et al., 2010; Simmons et al., 2010;
Umamaheswari et al., 2010).
2.3.2. 14M_zTDP2cat fluorescence-based biochemical assay
2.3.2.1. Optimization of assay conditions. All experiments were
performed in 384-well costar black plates. The reaction buffer used
was composed by 50 mM Tris-HCl pH 7.4, 10 mM MgCl2, 80 mM KCl,
0.05% (v/v) Tween-20, and 1 mM dithiothreitol (DTT). To optimize
enzyme and substrate concentrations, experiments varying the
concentration of the enzyme (0–800 pM) as well as experiments
varying substrate concentrations (0–50 nM) with fixed concentrations
of substrate (1 μM) and enzyme (6.25 pM) respectively were performed.
All experiments were performed with a final reaction volume of 20 μL:
To 10 μL of enzyme solution, 10 μL of substrate was added to initiate
reaction. The fluorescence of the product was measured using a
SpectraMax M5e (Molecular Devices) (λex 285 nm; λem 325 nm; λcutoff
315 nm) in kinetic-mode at 25 °C for 60–120 min. Assay DMSO
tolerance was tested using 10 concentrations of DMSO (0–10%), and
IC50 determination of the control (2) was performed by using 12 serial
(2-fold) dilution concentrations (0.92–2000 μM). Three consecutive
days experiments were also performed to determine the basic
“screenability” of the optimized conditions in the HTS format
employed.
2.3. Biology
Compound library of 1600 compounds was purchased from
Chembridge. Stock solutions between 2 and 20 mM were initially pre-
pared and then sets of 320 compounds were plated and stored into 384-
well plates at 2 mM in DMSO, mimicking the final HTS layout. Before
screening, an intermediate plate was prepared consisting of compounds
at 100 μM in reaction buffer. Substrate oligonucleotides 5′-(6-FAM-
NHS)(5′-tyrosine)GATCT(3′-BHQ-1)-3′, 5′-phosphotyrosine-GATCTAA
AAGACT-3′-(6-FAM)-3′ and 5′-phosphotyrosine-GATCTA(Cy5)AAGA
CTT-phosphate-3′ were purchased from Midland Certified Reagents.
2.3.2.2. Screening of the compound library. Compounds were screened at
single concentration of 50 μM (final DMSO concentration of 2.5%) in
two different plates. In each plate, maximum (wells containing enzyme,
substrate and vehicle DMSO, 32 wells) and minimum controls (without
enzyme, 24 wells) were included. To a black 384-well plate, 10 μL of
compound solution (100 μM) was added, followed by addition of 5 μL of
enzyme (25 pM, final concentration of 6.25 pM). After a pre-incubation
period
of
10 min,
5 μL
of
substrate
5′-(6-FAM-NHS)(5′-
2.3.1. Protein purification, crystallization and structure determination
The humanized zebrafish TDP2 catalytic domain (spanning amino
acid residues 120 to 369; 14M_zTdp2cat) was expressed from a codon-
optimized gene as a SUMO (Smt3)-fusion protein in the Escherichia coli
strain BL21(DE3) and purified essentially as described (Shi et al., 2012)
using nickel-affinity and size-exclusion chromatography steps. The first
residue Asp120 was mutated to Ser to facilitate cleavage by the SUMO-
protease Ulp1. The purified protein in 20 mM Tris-HCl (pH 7.4),
150 mM NaCl, 10 mM dithiothreitol was concentrated by ultrafiltration
to ~15 mg/mL, flash frozen in liquid nitrogen, and stored in −80 °C.
Human TDP2 catalytic domain (residues 110–362) was expressed and
purified similarly, except that SUMO fused to the N-terminus was kept
uncleaved to mitigate the lower solubility of the human enzyme.
14M_zTdp2cat was crystallized by the hanging drop vapor diffusion
method using a reservoir solution consisting of 22% polyethylene glycol
3350, 75 mM sodium malonate (pH 7.0), and 0.2 M NaCl. The crystals
typically grew as clusters of small and thin plates resembling a pine-
cone, and they had to be broken into small pieces before being flash
tyrosine)GATCT(3′-BHQ-1)-3′ (4 μM, final concentration of 1 μM) was
added, and the reaction was allowed to proceed at 25 °C for 60 min
before being stopped by the addition of 5 μL ethylenediaminetetraacetic
acid (EDTA) 0.5 mM, pH = 8. The fluorescence was measured using a
SpectraMax M5e (Molecular Devices) (λex 285 nm; λem 325 nm; λcutoff
315 nm) at 25 °C. Compounds were selected as hits in the HTS if they
displayed at least 24% enzyme inhibition. In this analysis, two
compounds exhibited a percent inhibition lower than −50% (TDP2
activity higher than 150%) and were excluded. In addition, due to
solubility issues in DMSO, ten compounds were tested at concentrations
lower than 50 μM.
For follow-up experiments, including IC50 determinations, the same
protocol was adopted, with the exception that the plates were analyzed
in kinetic-mode, and the fluorescence monitored constantly for 60 min
immediately after the reaction was started. When solubility allowed,
hits were further analyzed at four concentrations in triplicate (25, 50,
100, and 200 μM), to test dose-dependent response and their potential
interference with the assay. When 50% inhibition was reached up to
70

C.J.A. Ribeiro et al.
EuropeanJournalofPharmaceuticalSciences118(2018)67–79
200 μM, IC50s were determined when possible.
inhibitor solution (in DMSO, concentration 20-fold higher than the
tested concentration) or DMSO was added and allowed to mix for
10 min. Then, 2 μL of 5′-YGATCTAAAAGACT-3′-(6-FAM)-3′ substrate
(10 μM) was added, and the reaction was allowed to proceed at 25 °C
for 30 min before being stopped by the addition of 10 μL of formamide/
bromophenol blue loading buffer. Samples were boiled for 5 min,
loaded into 15% polyacrylamide TBE + urea gel, and ran at 200 V for
45 min. Gels were then analyzed by BioRad Gel Doc EZ imager (blue
tray 1708273).
2.3.2.3. Data analysis. Throughout the assay optimization and
validation steps the conditions tested were assessed by statistical
parameters: Z′ factor, signal-to-background (S:B) and coefficients of
variation (%CV) (Inglese et al., 2007; Zhang et al., 1999). These
parameters were also calculated during the HTS screen phase to
ensure the quality of the data obtained.
′
Z = 1 − (3 × σmax + 3 × σmin )/(μmax − μmin
)
S: B = μmax /μmin
%CV = σ/μ × 100
2.3.4.2. WT_zTDP2 and mutants performance evaluation. TDP2 reactions
were carried out similarly to describe previously (Gao et al., 2012) with
the following modifications. A 14-mer single-stranded oligonucleotide
DNA
substrate
5′-phosphotyrosine-GATCTA(Cy5)AAGACTT-3′-
In the equations above, μmax and μmin represent the mean values of
fluorescence for maximum and minimum controls, respectively, and
σmax and σmin represent their respective standard deviations.
phosphate containing a 5′-phosphotyrosine (Cy5-Y14) and Cyanine5
(Cy5) dye incorporated in place of the nucleotide at position 7. The
Cy5-Y14 was incubated at 40 nM with TDP2 in the absence or presence
of an inhibitor for 15 min at room temperature in the reaction buffer
containing 50 mM Tris-HCl, pH 7.5, 80 mM KCl, 5 mM MgCl2, 0.1 mM
EDTA, 1 mM DTT, 40 μg/mL BSA, and 0.01% Tween 20. Reactions were
terminated by the addition of 1 volume of gel loading buffer (5 mM
EDTA solution in formamide). Polyacrylamide gels were cast in low
fluorescence glass molds. Samples were subjected to a 16% denaturing
PAGE. Gels were scanned directly in the mold on a Typhoon™ FLA 9500
(GE Healthcare) scanner using 635 nm excitation light. Densitometry
analyses were performed using the ImageQuant software (GE
Healthcare).
Kinetic parameters (Km and Vmax) were calculated using a non-
linear Michaelis-Menten fitting of initial velocities as a function of
substrate concentration (GraphPad Prism software). The inhibitory ac-
tivity of the test compounds was calculated by normalizing the results
plate-wise against the controls according to the following equation:
Inhibition% = [1 − (Fc − μmin )/(μmax − μmin )] × 100
In the equation above, Fc is the fluorescence value from an in-
dividual test compound well, or in the case of dose-response experi-
ments, the mean value of three technical replicates performed in the
same plate. Data from the IC50 determination experiments using 10 to
12 concentrations of inhibitor and vehicle-alone were fitted by
GraphPad Prism software.
3. Results and discussion
2.3.3. SUMO hTDP2cat fluorescence-based biochemical assay
3.1. Chemistry
The reaction buffer used was composed by 50 mM Tris-HCl pH 7.4,
10 mM MgCl2, 80 mM KCl, 0.05% (v/v) Tween-20, and 1 mM DTT. To a
black 384-well plate, 10 μL of compound solution (in reaction buffer,
concentration 2-fold higher than the tested concentration) was added,
followed by addition of 5 μL of SUMO hTDP2cat enzyme (12.5 pM, final
concentration of 3.13 pM). After a pre-incubation period of 10 min, 5 μL
of substrate 5′-(6-FAM-NHS)(5′-tyrosine)GATCT(3′-BHQ-1)-3′ (1 μM,
final concentration of 0.25 μM) was added, and the reaction was al-
lowed to proceed at 25 °C for 60 min. The fluorescence was measured,
and the data was processed as described previously for 14M_zTDP2.
Substituted 1,2-dihydro-2,2,4-trimethylquinoline derivatives (10,
Scheme 1) (Li et al., 2015) were synthesized via modified Skraup re-
action by treating substituted anilines with acetone in the presence of
catalytic InCl3. Acid catalyzed reaction of compound 10 with cyano-
guanidine produced the desired 4-methylquinazolyl-2-guanidines ana-
logs (11) (Rosowsky and Modest, 1972; Webb et al., 2003) which was
further derivatized by treating with diketoester or anthranilic acid in
DMF at 170 °C to furnish compounds 12a (LaPorte et al., 2014) and 12b
(Shikhaliev et al., 2003) respectively. Compound 15 was obtained by
the condensation of anthranilic acid derivative (13) with cyanoguani-
dine in refluxing 10% sulfuric acid in a moderate yield.
2.3.4. TDP2 gel-based assays.
2.3.4.1. Confirmatory assay. The reaction buffer used was composed by
50 mM Tris-HCl pH 7.4, 10 mM MgCl2, 80 mM KCl, and 1 mM DTT. To
6.5 μL of reaction buffer, 1 μL of enzyme solution (15 μM), and 0.5 μL of
Reaction of furoic hydrazide (16, Scheme 2) with carbon disulfide in
methanolic potassium hydroxide at room temperature afforded the di-
thiocarbazinate derivative (17) which subsequently underwent ring
Scheme 1. Synthesis of N-quinazolin-2yl guanidines. (a) InCl3 (5 mol%) 50 °C, 24 h, 40–65%; (b) cyanoguanidine, HCl conc, H2O, reflux, 4 h, 55–62%; (c) ethyl 2-methylacetoacetate,
DMSO, 170 °C, 30%; (d) Anthranilic acid, DMSO, 170 °C, 64%; (e) H2SO4 conc, 100 °C, 55%.
71

C.J.A. Ribeiro et al.
EuropeanJournalofPharmaceuticalSciences118(2018)67–79
Scheme 2. Synthesis of triazolo[3,4-b]
[1,3,4]thiadiazoles. (a) CS2, KOH, MeOH, r.t,
24 h; (b) N2H4·H2O, reflux, 5 h, 70% over
two steps; (c) RCOOH, POCl3, reflux, 6 h,
53–70%.
Fig. 2. The humanized 14M_zTDP2 is catalytically as active as the WT enzyme. (A) Hydrolysis of 5′-phosphotyrosyl DNA substrate catalyzed by WT_zTDP2, and the 14 M, 6 M, 10 M
mutants. (A1) Representative gels: lanes 1, 9, 18, 26: substrate DNA alone; lanes 2–8: seven 2-fold dilutions of WT_zTDP2 from 3.2 nM to 50 pM; lanes 10–17: seven 2-fold dilutions of
14M_zTDP2 mutant from 3.2 nM to 50 pM; lanes 19–25: seven 2-fold dilutions of 6M_zTDP2 mutant from 400 nM to 6.25 nM; lanes 27–33: seven 2-fold dilutions of 10M_zTDP2 mutant
from 400 nM to 6.25 pM. (A2) 5′-Phosphotyrosyl DNA substrate conversion quantification and regression curves, plotted data and SD are based on three independent repeats. (B)
Inhibition of zTDP2 (WT and 14 M mutant) catalyzed hydrolysis of 5′-phosphotyrosyl DNA substrate by deazaflavin 2. (B1) Representative gels: lane 1: substrate DNA alone; lane 2:
substrate DNA with 14M_zTDP2; lanes 3–14: substrate DNA, 14M_zTDP2, and twelve 3-fold dilutions of 2 from 111 μM to 0.63 nM; lanes 15–22: substrate DNA, 14M_zTDP2, eight 3-fold
dilutions of 2 from 111 μM to 50 nM. (B2) 5′-Phosphotyrosyl DNA hydrolysis inhibition quantification and regression curves.
closure with an excess of hydrazine monohydrate to give the key tria-
zole intermediate (18) (Khan et al., 2016). Condensation of the aryl
triazole (18) with substituted benzoic acids in phosphorus oxychloride
under reflux provided access to the 1,2,4-triazolo[3,4-b][1,3,4]thia-
diazoles (19) in moderate to excellent yields.
2016). Thus, to obtain crystal structures relevant for the development
of inhibitors against human TDP2, we took the approach to selectively
humanize the active site and the DNA-binding groove of zTDP2 by
amino acid substitutions and to make a surrogate for hTDP2 with
adequate biochemical properties. Recently, Hornyak et al. similarly
generated a humanized mouse TDP2 to overcome the problem of poor
quality and reproducibility of the hTDP2cat crystals (Hornyak et al.,
2016). We also took this approach to make mouse TDP2 sensitive to
deazaflavin inhibitors (Marchand et al., 2016).
To achieve our goal of generating humanized zebrafish TDP2 pro-
teins, several surface-exposed residues within ~20 Å from the active
center were changed to the corresponding human sequence. A series of
zTDP2 catalytic domain constructs (residues 120–369) were generated.
The three reported here included 6, 10 or 14 amino acid substitutions:
6 M (L136N, A139S, C239T, K240R, K314L, C322L), 10 M (L136N,
A139S, C239T, K240R, K314L, Y318A, V319A, S320C, R321K, C322L)
3.2. Development of a protein surrogate: 14M_zTDP2cat
Previous attempts to obtain high-quality crystals of hTDP2 were
unsuccessful due to limited solubility of the human protein. Conversely,
zTDP2 is a much more soluble protein with suitable crystallographic
behavior and that has provided structures with high resolution (Rao
et al., 2016; Shi et al., 2012). In addition, the catalytic domains of
zebrafish and human TDP2 share high sequence identity (58%) war-
ranting high degree of structural similarity (Shi et al., 2012). Yet, they
do not respond similarly to the deazaflavin inhibitors (Marchand et al.,
72

C.J.A. Ribeiro et al.
EuropeanJournalofPharmaceuticalSciences118(2018)67–79
Fig. 3. A) The active site of 14M_zTDP2.
2mFo-DFc electron density map contoured
at 1.0σ is overlaid on the atomic model of
the protein shown in sticks. Magnesium ion
and coordinating water molecules are re-
presented by small green and red spheres,
respectively, with hydrogen bonds or metal-
coordinating interactions indicated by
yellow dotted lines. Other water molecules
are depicted by red crosses. Protein residues
are labeled accordingly to the wild type
zebrafish TDP2 sequence, with the num-
bering for the equivalent human TDP2 re-
sidues
shown
in
parentheses.
B)
Superposition of hTDP2 (pdb code: 5J3P)
and 14M_zTDP2 (pdb code: 6CA4) in orange
and green, respectively, with key active site
residues in sticks and labeled as follows:
hTDP2 (black text on white background),
14M_zTDP2 (white text on grey back-
ground). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
and 14 M (L136N, A139S, C239T, K240R, K309Q, K314L, T315G,
V316I, P317T, Y318A, V319A, S320C, R321K, and C322L). To survey
the performance of the WT_zTDP2 along with its generated 6 M, 10 M
and 14 M mutants, we redesigned the assay based on use of 32P-labeled
DNA (Cortes Ledesma et al., 2009; Gao et al., 2012) to employ the
fluorophore marker for oligomeric DNA detection. Using this gel-based
assay with Cyanine5 (Cy5) dye incorporated into DNA substrate allows
for speedier assays by eliminating required substrate 3′-32P labeling
reaction as well as cutting the gel handling steps, i.e. transferring from
the mold, drying and exposing. With this assay we verified that only
14M_zTDP2 is catalytically as active as WT_zTDP2 (Fig. 2A), while 6 M
and 10 M mutants turned out to be significantly less active (Fig. 2A). In
addition, 14M_zTDP2 is as sensitive as hTDP2 to the known hTDP2
inhibitor 2: hTDP2 IC50 = 40 nM (Marchand et al., 2016; Raoof et al.,
2013); 14M_zTDP2 IC50 = 63 nM; zTDP2 IC50 > 100 μM (Fig. 2B).
We were able to obtain high-quality crystals of 14M_zTDP2, which
allowed structure solution at 1.62 Å resolution (Rfree = 20.9%). The
crystal structure of our humanized TDP2 (14M_zTDP2) shows the active
site containing a magnesium ion closely resembling those of mouse and
zebrafish TDP2 (Fig. 3A). Notably, an extended loop flanking the active
site of 14M_zTDP2 appears to be highly flexible, where the residues
312–318 (corresponding to hTDP2 303–309) could not be modeled due
to weak electron density. This flexible loop was previously shown to
play roles in the binding of DNA (Schellenberg et al., 2012; Shi et al.,
2012) and deazaflavin-based inhibitors (Hornyak et al., 2016). A su-
perposition of our structure of 14M_zTDP2 refined at 1.6 Å resolution
with that of hTDP2cat (3.1 Å resolution) shows high structural similarity
as expected, with an r.m.s.d. of 0.47 Å for all backbone atoms (Fig. 3B).
We used this surrogate protein for in silico and biochemical studies.
Fig. 4. Schematic representation of the in silico screening campaign. vHTS: virtual high
throughput screening; SP: standard precision; XP: extra-precision.
3.3. Curation of screening compounds
The protein model used for the VHTS was based on the crystal
structure of the humanized 14M_zTDP2 (6CA4). 460,000 compounds
from Chembridge Library were then run in this model using Glide vHTS
and scored based on active site binding. The top 7200 compounds were
selected for visual inspection which concerns mostly physicochemical
properties and synthetic tractability of ranked compounds. In the end,
1600 small molecules were purchased (Fig. 4).
3.4. Assay optimization
Fig. 5. Schematic representation of the biochemical fluorescence-based assay, employing
We developed
a fluorescence based assay suitable for high
a
fluorophore-quencher dual-labeled substrate. Upon cleavage by humanized
throughput screening in 384-well plates. The substrate employed con-
sists of a fluorophore-quencher dual-labeled single stranded 5-mer oli-
gonucleotide (Fig. 5). Specifically, it contains a 5′-tyrosine residue la-
beled with the donor fluorophore FAM and a 3′-BHQ-1 molecule that
14 M_zTDP2cat enzyme, labeled tyrosine is released from the single-stranded oligonu-
cleotide. TDP2 activity is monitored by the increase of fluorescence measured at 525 nm,
when FAM is excited at 485 nm. FAM: 6-Carboxyfluorescein; BHQ-1: Black hole
quencher-1.
73

C.J.A. Ribeiro et al.
EuropeanJournalofPharmaceuticalSciences118(2018)67–79
Fig. 6. HTS optimization. (A) Enzyme titration as a function of substrate conversion over reaction time. [S] = 1 μM; (B) Michaelis-Menten curve for 14M_zTDP2cat. [E] = 6.25 pM; (C)
Linearity of enzymatic response, when [E] = 6.25 pM and [S] = 1 μM. R2 = 1; Substrate conversion after 60 min: 10.1
1.8%; S:B = 10 (three independent experiments performed in
triplicate); (D) Dose-response curve for the positive control 2. Km, Vmax, and IC50 values are the mean of three independent experiments performed in triplicate, with error bars
representing one standard deviation. RFU: Relative fluorescence units.
acts as a dark quencher. The close proximity between FAM and BHQ-1
in the intact substrate allows proper quenching of the energy emitted at
525 nm by FAM when excited at 485 nm. Cleavage of the phospho-
diester 5′-phosphotyrosine bond mediated by TDP2 liberates the
fluorophore from the quencher effect, and an increase in fluorescence is
observed. Therefore, TDP2 activity can be monitored as a function of
the fluorescence measured.
60 min. This assay allowed us to determine a starting enzyme con-
centration (6.25 pM) for subsequent kinetic studies that ensured line-
arity of the fluorescence measured over the duration of the assay
(Fig. 6A) To choose the appropriate substrate concentration, we then
performed experiments with different substrate amounts (0–50 μM)
with the enzyme concentration selected in the preceding step to de-
termine the Michaelis-Menten steady-state kinetic parameters (Km and
Vmax) (Fig. 6B). We decided to maintain the sub-Km substrate con-
centration tested previously (1 μM, 0.25 × Km) that also allowed a
We started the optimization by varying protein concentration
(0–800 pM) with a fixed amount of substrate (1 μM) over a period of
Fig. 7. HTS assays. (A) Example of one plate reading, showing only the maximum (Max, enzyme + substrate) and minimum (Min, substrate only) and their respective average value (μ)
and 3 × Standard deviation (3σ) limits. Statistical parameters: Z′ = 0.78; S:B = 10; intra-plate %CVMax = 6% and %CVMin = 5%. Data were plotted first by column, then by row. RFU:
Relative fluorescence units (B) Scatter plot showing the percent inhibition for the compound library.
74

C.J.A. Ribeiro et al.
EuropeanJournalofPharmaceuticalSciences118(2018)67–79
Table 1
Good assay reproducibility and robustness is characterized by high
Z′-factors (between 0.5 and 1), low coefficients of variance (lower than
15%) and consistent signal to background (S:B) windows (Inglese et al.,
2007; Zhang et al., 1999). Three consecutive days experiments mi-
micking the HTS conditions revealed a 0.72 ≤ Z′ ≤ 0.81, intra-plate
and inter-plate CV < 10%, and S:B of 9–10.
Screening hits that showed ≥24% inhibition in the screening, and at least 50% inhibition
at 200 μM inhibitor concentration.
Compound hits
IC50 (μM)a
Compound hits
IC50 (μM)a
20
21
18.2
26.5
33.1
46.3
1.7
27
11a
28
29
30
31
32
33
100 μM: 50
5%b
0.3
3.2
9.3
129
133
2
11
19a
22a
23
22b
24
3.5. HTS and preliminary SAR
150 μM: 51
150 μM: 50
200 μM: 56
200 μM: 55
200 μM: 50
2%b
1%b
5%b
5%b
1%b
50 μM: 50
5%b
63.2
83.8
85.2
92.0
0.3
0.9
5.4
14.1
A
library of 1600 compounds previously curated via virtual
screening was screened at a single concentration of 50 μM (Fig. 7). Z′-
factors between 0.72 and 0.84 were obtained for all plates, and daily
intra-plate and inter-plate CV < 10%. From this library screening, 32
compounds (2% of the library) were selected for further studies (in-
hibition ≥24%). First, preliminary studies consisting in experiments
involving four compound concentrations (25, 50, 100, and 200 μM,
when solubility allowed) were performed, and then for hits with asso-
ciated dose-dependent inhibition, IC50 values were calculated (Table 1,
Fig. 8). To obtain potentially better lead compounds, we eliminated all
hits with a steep dose-response curve with slopes > 2 (22b, 24, 25,
Fig. 8A), and compounds containing moieties associated with pan assay
interference compounds (PAINS). These types of structure are linked to
promiscuous bioactivity due to their reactivity nature rather than de-
sired enzyme inhibition (20, 26, 28, Fig. 8B) (Baell and Holloway,
2010; Dahlin and Walters, 2016).
To decide which molecules to move forward to preliminary SAR
studies (Fig. 8C), we analyzed their physicochemical properties and
determined a few efficiency parameters: ligand efficiency (LE) (Hopkins
et al., 2004), lipophilic ligand efficiency (LLE) (Leeson and
Springthorpe, 2007; Shultz, 2013), and ligand efficiency dependent li-
pophilicity (LELP) (Keseru and Makara, 2009) (Table 2). Since TDP2
activities were in the two-three digits micromolar activity range, we
25
26
a,bValues are the means of at least two independent experiments performed in triplicate.
a
Compounds listed according to their IC50 values (chemical structures are presented in
Fig. 8).
b
When solubility limitations in the assay conditions prevented accurate IC50 values to
be determined, the first concentration tested that reached 50% inhibition is presented,
and their chemical structures are reported in Fig. S1 (Supplementary material).
good signal-to-background ratio (10:1), with a low substrate conversion
(10.1
1.8%) (Fig. 6C). In addition, under the established conditions,
the assay showed a DMSO tolerance up to 2.5% without significant
signal loss (fluorescence variability < 5% from control without DMSO).
To verify the accuracy of the inhibition measurements and to confirm
the adequate surrogacy of 14M_TDP2cat protein, we determined the IC50
of deazaflavin 2, which has been extensively studied in different bio-
chemical assays (NPPP chromogenic assay IC50 = 40 nM (Raoof et al.,
2013); gel-based assay IC50 = 40
fluorescence-based assay IC50 = 19
We obtained an IC50 of 40.5
literature (Fig. 6D).
3 nM (Marchand et al., 2016);
3.2 nM (Hornyak et al., 2016)).
2.7 nM which is in accordance with
Fig. 8. (A) Compounds with a dose-response slope > 2; (B) Pan assay interference compounds (PAINS); (C) Hit compounds selected for further evaluation.
75

C.J.A. Ribeiro et al.
EuropeanJournalofPharmaceuticalSciences118(2018)67–79
Table 2
Physicochemical properties and efficiency parameters of selected hits.
Compds
RBa
cLogPa
tPSAa
Hacca
Hdona
MWa
LLEa,b
LEa,c
LELPa,d
11a
19a
21
2
4
3
4
1.58
4.19
3.31
4.96
96.91
56.22
115.3
78.39
4
4
7
6
3
0
1
1
231.26
296.35
360.33
378.79
2.3
0.32
0.30
0.24
0.22
4.9
14
14
22
0.29
1.3
22a
–0.62
aRB: rotatable bonds; cLogP: logarithm of its partition coefficient between n-octanol and water log(coctanol/cwater); tPSA: topological polar surface area; Hacc: Hydrogen bond acceptor;
Hdon: Hydrogen bond donner; MW: molecular weight; LLE: lipophilic ligand efficiency, also referred as LipE; LE: ligand efficiency; LELP: ligand efficiency dependent lipophilicity.
bLLE = pIC50 − cLogP.
cLE = 1.4 × pIC50 / HAC; HAC: heavy atom count (i.e., number of non-hydrogen atoms).
dLELP = LE / cLogP.
Non-violations to the rule of three and to the efficiency parameters established are highlighted in grey.
decided to increase the parameters constrains to the Lipinski rule-of-
five (Lipinski et al., 2001) to match with what it is typically observed in
fragment-based drug design libraries, such as the rule-of-three:
MW < 300 Da, normally corresponding to < 20 non‑hydrogen atoms;
≤3 hydrogen-bond donors and acceptors, ≤3 rotatable bonds,
cLogP ≤ 3, and polar surface area (PSA) ≤ 60 Å (Congreve et al., 2003;
Jhoti et al., 2013). In addition, a good hit in early optimization stages
should detain a LLE and a LE of at least 2 and 0.3, respectively, with a
LELP ≤ 10 (Keseru and Makara, 2009; Leach et al., 2006). The HTS hit
closer to fulfill these requisites was compound 11a that showed better
efficiency parameters despite its lower potency, due presumably to its
lower clogP and molecular weight values. The second compound that
we selected for further analysis was compound 19a that showed the
second-best features in both efficiency parameters and rule-of-three
violations.
human TDP2cat. As observed in Fig. 9 both hits were able to inhibit the
hTDP2 in the same extent as in the primary assay, further corroborating
the good surrogacy of 14M_zTDP2 enzyme.
3.5.1. Triazolo[3,4-b][1,3,4]thiadiazole derivatives
Recently new biological activities were described for 3,6-diaryl and
heteroaryl derivatives of triazolo[3,4-b][1,3,4]thiadiazole derivatives
(19) such as antitubulin activity (Xu et al., 2017), dCTP pyropho-
sphatase 1 inhibition (Llona-Minguez et al., 2017), CYP1A1 inhibition
(Joshi et al., 2017), and anti-angiogenesis activity by inhibiting TIE-2
(Pan et al., 2015). The observed TDP2 inhibitory activity of hit 19a was
confirmed through resynthesizing and retesting (Table 3). This enable
us to test an additional 41 triazolo[3,4-b][1,3,4]thiadiazole compounds
to complement the 24 analogues already present in the 1600 compound
library (Table S2, supplementary material). The combined set of 65
compounds was screened at 100 μM. IC50 determination was conducted
for derivatives with > 50% inhibition at this concentration (19c–e). We
To confirm the TDP2 inhibition potential of these two hits, we also
designed an orthogonal fluorescence gel-based assay that employs
Fig. 9. (A) Schematic representation of the fluorescence gel-based assay. Upon cleavage by SUMO-hTDP2cat enzyme, tyrosine is cleaved from the single-stranded oligonucleotide. Both
oligonucleotide substrate and product detain the 3′-fluorophore. FAM: 6-Carboxyfluorescein; (B) Representative gel: lanes 1, 10: substrate DNA only; lane 2: maximum response; lane 3:
positive control 2 at IC50 concentration; lanes 4–6: three 10-fold dilutions of hit compound 11a; lanes 7–9: three 10-fold dilutions of hit compound 19a; S: substrate; P: product.
Table 3
6,7-Dihydrobenzo[d][1,2,4]triazino[6,5-f][1,3]oxazepane derivatives.
Compounda
19a
R1
R2
IC50 (µM)b
20.7 1.3
> 100
3,4–dimethylphenyl
2–methoxyphenyl
2–methoxyphenyl
2,4–dichlorophenyl
2–methoxyphenyl
furan–2–yl
furan–2–yl
3–methylphenyl
Pyridin–2–yl
2–bromophenyl
19b
19c
17.6 0.8
31.5 1.6
74.9 4.2
19d
19e
76

C.J.A. Ribeiro et al.
EuropeanJournalofPharmaceuticalSciences118(2018)67–79
Table 4
N-quinazolin-2yl guanidines derivatives.
Compounda
IC50 (µM)b
R1
R2
R3
6-methoxy
H
-
114
8
11a
11b
150 µM inhibits
50.9 1.1%c
6-ethoxy
6-chloro
H
H
-
-
>200
>200
>200
>100
>100
>100
11c
11d
15
6,7-dimethyl Benzoyl
-
-
-
-
H
H
H
H
H
34a
34b
34c
6-methoxy
6-methyl
Methyl
Methyl
25 µM inhibits
58.7 8.1%c
6-methoxy
2-methoxyphenyl
H
12c
8-ethoxy
Methyl
Methyl
ethyl
19.1 0.3
22.8 1.9
27.1 0.3
12d
12a
12b
6-methoxy
6-methoxy
methyl
Phenyl
aKnown bioactivities for the synthesized compounds against other targets are reported in Table S3 (Supplementary material).
bValues are the means of three independent experiments performed in triplicate.
cWhen solubility limitations in the assay conditions prevented accurate IC50 values to be determined, the first concentration tested that reached 50% inhibition is presented.
also synthesized derivative 19b, which was designed via hybridizing
19a and 19c/19e. Yet, 19b did not exhibit significant TDP2 inhibition
at concentrations up to 100 μM. Overall, these preliminary SAR efforts
identified a total of 4 derivatives with IC50 lower than 100 μM
(Table 3).
Table 5
Comparison of IC50 values when 14M_zTDP2 and hTDP2 enzymes were employed.
Compound
14M_zTDP2cat IC50 (μM)a
SUMO hTDP2cat IC50 (μM)a
11a
12a
19a
114
8
113
10
1.7
1.9
aKnown bioactivities for the synthesized compounds against other
targets are reported in Table S3 (Supplementary material).
bValues are the means of three independent experiments performed
in triplicate.
22.8
20.7
1.9
1.3
17.7
23.6
a
Values are the means of three independent experiments performed in triplicate.
3.5.2. 2-Guanidine-4-methylquinazoline derivatives
polymerase III inhibition (Guiles et al., 2009). We purchased and tested
two derivatives 12c–d. Both were at least 5-fold more active than the
guanidine hit. This led to the synthesis and testing of two new pyr-
imidone derivatives (12a–b), with the 6-methoxy group of the original
hit retained. All pyrimidone derivatives tested were found to be equi-
potent (IC50 = 17–27 μM).
Lastly, the activity of the hit compounds, as well as the best syn-
thesized guanidine derivative, was confirmed using human TDP2 en-
zyme in the same biochemical fluorescence-based assay, as similar
potencies were observed (Table 5).
2-Guanidine-4-methylquinazoline analogues are associated with a
range of biological activities, such as heme enzyme myeloperoxidase
inhibition (Soubhye et al., 2017), antibacterial activity by inhibiting
translation (Komarova et al., 2017), A type γ-aminobutyric acid re-
ceptors (GABAAR) antagonism (Xiao et al., 2013), acid sensing ion
channel ASIC3 activation (Yu et al., 2010), and anti-H2-histamine ac-
tivity (Pinelli et al., 1996). Again, hit 11a was confirmed via re-
synthesizing and retesting (Table 4). The initial SAR study focusing on
position 6 (11a–c) revealed that increasing the length of the ether side
chain did not interfere significantly with the activity (11b vs. 11a),
whereas changing the substituent to chloro abrogated completely the
inhibitory activity up to 200 μM (11c vs. 11a). In addition, no inhibi-
tion was observed with the 4-ketone analogue (15) synthesized to probe
position 4 of the hit compound. Four compounds with derivatization in
the guanidine end were already present in the screening library (11d,
34a–c). However, pyrimidine derivatives (34a–c) were not able to
produce 50% inhibition at 100 μM and higher concentrations were not
achievable due to solubility limitations. Although aromatization of the
guanidine did not lead to potency improvement, we decided to test
pyrimidone derivatives due to their reported biological activities, such
as STAT3 pathway inhibition (LaPorte et al., 2014), enhancement of
zeste 2 (EZH2) inhibition (Wu et al., 2016), and bacterial DNA
4. Conclusions
TDP2 inhibitors can be used to enhance TOP2-targeted drugs (eto-
poside, doxorubicin, daunorubicin), which are widely used as antic-
ancer therapies. In addition, TDP2 inhibitors could provide novel an-
tiviral strategies for targeting picornaviruses and HBV. To facilitate the
discovery of novel TDP2 inhibitor scaffolds, we report here the gen-
eration of a humanized zebrafish TDP2 (14M_zTDP2), which provides
further structural insights on the TDP2 catalytic site. We show that the
14 M enzyme is sensitive to the reference TDP2 inhibitors of the dea-
zaflavin class, and is readily amenable to biochemical and drug
screening studies. Indeed, we developed a new fluorescence-based
77

C.J.A. Ribeiro et al.
EuropeanJournalofPharmaceuticalSciences118(2018)67–79
biochemical assay suitable for 384-well plate screenings with a simple,
homogeneous “mix and read” procedure that allows readings of both
continuous and quenched reactions. This assay was used to screen 1600
commercial compounds preselected from 460,000 Cambridge Library
compounds through vHTS. Selected hits (11a and 19a) were confirmed
Friesner, R.A., Banks, J.L., Murphy, R.B., Halgren, T.A., Klicic, J.J., Mainz, D.T., Repasky,
M.P., Knoll, E.H., Shelley, M., Perry, J.K., Shaw, D.E., Francis, P., Shenkin, P.S., 2004.
Glide: a new approach for rapid, accurate docking and scoring. 1. Method and as-
sessment of docking accuracy. J. Med. Chem. 47, 1739–1749.
Friesner, R.A., Murphy, R.B., Repasky, M.P., Frye, L.L., Greenwood, J.R., Halgren, T.A.,
Sanschagrin, P.C., Mainz, D.T., 2006. Extra precision glide: docking and scoring in-
corporating a model of hydrophobic enclosure for protein-ligand complexes. J. Med.
Chem. 49, 6177–6196.
using
a novel secondary gel-based fluorescence assay with re-
synthesized samples. Preliminary SAR studies identified guanidine de-
rivative 12a (hTDP2 IC50 = 17.7 μM) as an improved inhibitor over the
original hit 11a (hTDP2 IC50 = 113 μM). These results support our new
assays (chemical screening and crystallographic analyses) as tools in
TDP2 inhibitor discovery.
Gao, R., Huang, S.Y., Marchand, C., Pommier, Y., 2012. Biochemical characterization of
human tyrosyl-DNA phosphodiesterase 2 (TDP2/TTRAP): a Mg(2+)/Mn(2+)-de-
pendent phosphodiesterase specific for the repair of topoisomerase cleavage com-
plexes. J. Biol. Chem. 287, 30842–30852.
Gomez-Herreros, F., Romero-Granados, R., Zeng, Z., Alvarez-Quilon, A., Quintero, C., Ju,
L., Umans, L., Vermeire, L., Huylebroeck, D., Caldecott, K.W., Cortes-Ledesma, F.,
2013. TDP2-dependent non-homologous end-joining protects against topoisomerase
II-induced DNA breaks and genome instability in cells and in vivo. PLoS Genet. 9,
e1003226.
Acknowledgments
Guiles, J., Sun, X., Critchley, I.A., Ochsner, U., Tregay, M., Stone, K., Bertino, J., Green, L.,
Sabin, R., Dean, F., Dallmann, H.G., McHenry, C.S., Janjic, N., 2009. Quinazolin-2-
ylamino-quinazolin-4-ols as novel non-nucleoside inhibitors of bacterial DNA poly-
merase III. Bioorg. Med. Chem. Lett. 19, 800–802.
Hoa, N.N., Shimizu, T., Zhou, Z.W., Wang, Z.Q., Deshpande, R.A., Paull, T.T., Akter, S.,
Tsuda, M., Furuta, R., Tsutsui, K., Takeda, S., Sasanuma, H., 2016. Mre11 is essential
for the removal of lethal topoisomerase 2 covalent cleavage complexes. Mol. Cell 64,
580–592.
This research was supported by the Academic Health Center Faculty
Research Development Grant Program (FRD #14.23), University of
Minnesota, and partially by the Center for Drug Design, University of
Minnesota, and NIH grant GM118047 to HA. We thank Surajit Banerjee
for his assistance during diffraction data collection at the Advanced
Photon Source. This work is based upon research conducted at the
Northeastern Collaborative Access Team beamlines, which are funded
by the National Institute of General Medical Sciences from the National
Institutes of Health (P41 GM103403). This research used resources of
the Advanced Photon Source, a U.S. Department of Energy (DOE) Office
of Science User Facility operated for the DOE Office of Science by
Argonne National Laboratory under Contract No. DE-AC02-
06CH11357. Studies performed by EK, AG and YP were supported by
the Center for Cancer Research, the Intramural Program of the National
Cancer Institute (Z001 BC006150).
Hopkins, A.L., Groom, C.R., Alex, A., 2004. Ligand efficiency: a useful metric for lead
selection. Drug Discov. Today 9, 430–431.
Hornyak, P., Askwith, T., Walker, S., Komulainen, E., Paradowski, M., Pennicott, L.E.,
Bartlett, E.J., Brissett, N.C., Raoof, A., Watson, M., Jordan, A.M., Ogilvie, D.J., Ward,
S.E., Atack, J.R., Pearl, L.H., Caldecott, K.W., Oliver, A.W., 2016. Mode of action of
DNA-competitive small molecule inhibitors of tyrosyl DNA phosphodiesterase 2.
Biochem. J. 473, 1869–1879.
Inglese, J., Johnson, R.L., Simeonov, A., Xia, M., Zheng, W., Austin, C.P., Auld, D.S., 2007.
High-throughput screening assays for the identification of chemical probes. Nat.
Chem. Biol. 3, 466–479.
Jhoti, H., Williams, G., Rees, D.C., Murray, C.W., 2013. The ‘rule of three’ for fragment-
based drug discovery: where are we now? Nat. Rev. Drug Discov. 12 (644–644).
Joshi, P., McCann, G.J.P., Sonawane, V.R., Vishwakarma, R.A., Chaudhuri, B., Bharate,
S.B., 2017. Identification of potent and selective CYP1A1 inhibitors via combined
ligand and structure-based virtual screening and their in vitro validation in sac-
chrosomes and live human cells. J. Chem. Inf. Model. 57, 1309–1320.
Kabsch, W., 2010. XDS. Acta Cryst. D 66, 125–132.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.ejps.2018.03.021.
Kankanala, J., Marchand, C., Abdelmalak, M., Aihara, H., Pommier, Y., Wang, Z., 2016.
Isoquinoline-1,3-diones as selective inhibitors of tyrosyl DNA phosphodiesterase II
(TDP2). J. Med. Chem. 59, 2734–2746.
Keseru, G.M., Makara, G.M., 2009. The influence of lead discovery strategies on the
properties of drug candidates. Nat. Rev. Drug Discov. 8, 203–212.
Khan, I., Panini, P., Khan, S.U.-D., Rana, U.A., Andleeb, H., Chopra, D., Hameed, S.,
Simpson, J., 2016. Exploiting the role of molecular electrostatic potential, deforma-
tion density, topology, and energetics in the characterization of S⋯N and Cl⋯N su-
pramolecular motifs in crystalline triazolothiadiazoles. Crys. Growth Des. 16,
1371–1386.
Komarova, E.S., Osterman, I.A., Pletnev, P.I., Ivanenkov, Y.A., Majouga, A.G., Bogdanov,
A.A., Sergiev, P.V., 2017. 2-Guanidino-quinazolines as a novel class of translation
inhibitors. Biochimie 133, 45–55.
Koniger, C., Wingert, I., Marsmann, M., Rosler, C., Beck, J., Nassal, M., 2014. Involvement
of the host DNA-repair enzyme TDP2 in formation of the covalently closed circular
DNA persistence reservoir of hepatitis B viruses. Proc. Natl. Acad. Sci. U. S. A. 111,
E4244–E4253.
Kont, Y.S., Dutta, A., Mallisetty, A., Mathew, J., Minas, T., Kraus, C., Dhopeshwarkar, P.,
Kallakury, B., Mitra, S., Uren, A., Adhikari, S., 2016. Depletion of tyrosyl DNA
phosphodiesterase 2 activity enhances etoposide-mediated double-strand break for-
mation and cell killing. DNA Repair 43, 38–47.
References
Acker, M.G., Auld, D.S., 2014. Considerations for the design and reporting of enzyme
assays in high-throughput screening applications. Perspect. Sci. 1, 56–73.
Adams, P.D., Afonine, P.V., Bunkoczi, G., Chen, V.B., Davis, I.W., Echols, N., Headd, J.J.,
Hung, L.W., Kapral, G.J., Grosse-Kunstleve, R.W., McCoy, A.J., Moriarty, N.W.,
Oeffner, R., Read, R.J., Richardson, D.C., Richardson, J.S., Terwilliger, T.C., Zwart,
P.H., 2010. PHENIX: a comprehensive python-based system for macromolecular
structure solution. Acta Cryst. D 66, 213–221.
Adhikari, S., Karmahapatra, S.K., Elias, H., Dhopeshwarkar, P., Williams, R.S., Byers, S.,
Uren, A., Roy, R., 2011. Development of a novel assay for human tyrosyl DNA
phosphodiesterase 2. Anal. Biochem. 416, 112–116.
Baell, J.B., Holloway, G.A., 2010. New substructure filters for removal of pan assay in-
terference compounds (PAINS) from screening libraries and for their exclusion in
bioassays. J. Med. Chem. 53, 2719–2740.
Beck, D.E., Lv, W., Abdelmalak, M., Plescia, C.B., Agama, K., Marchand, C., Pommier, Y.,
Cushman, M., 2016. Synthesis and biological evaluation of new fluorinated and
chlorinated indenoisoquinoline topoisomerase I poisons. Bioorg. Med. Chem. 24,
1469–1479.
Brown, J.P., 1965. Reactions of 2,2-dialkyl-1,2-dihydroquinolines. I. Preparation of 2-
guanidinoquinazolines. Monsanto Tech. Revue 10, 16–18.
Cain, R., Narramore, S., McPhillie, M., Simmons, K., Fishwick, C.W.G., 2014. Applications
of structure-based design to antibacterial drug discovery. Bioorg. Chem. 55, 69–76.
Congreve, M., Carr, R., Murray, C., Jhoti, H., 2003. A ‘rule of three’ for fragment-based
lead discovery? Drug Discov. Today 8, 876–877.
Cortes Ledesma, F., El Khamisy, S.F., Zuma, M.C., Osborn, K., Caldecott, K.W., 2009. A
human 5′-tyrosyl DNA phosphodiesterase that repairs topoisomerase-mediated DNA
damage. Nature 461, 674–678.
Curtin, N.J., 2012. DNA repair dysregulation from cancer driver to therapeutic target.
Nat. Rev. Cancer 12, 801–817.
Koparır, M., Cansız, A., Demirdağ, A., 2004. Synthesis of some new 4,5-substituted-4H-
1,2,4-triazole-3-thiol derivatives. Molecules 9, 204.
Kossmann, B.R., Abdelmalak, M., Lopez, S., Tender, G., Yan, C., Pommier, Y., Marchand,
C., Ivanov, I., 2016. Discovery of selective inhibitors of tyrosyl-DNA phosphodies-
terase 2 by targeting the enzyme DNA-binding cleft. Bioorg. Med. Chem. Lett. 26,
3232–3236.
Laev, S.S., Salakhutdinov, N.F., Lavrik, O.I., 2016. Tyrosyl-DNA phosphodiesterase in-
hibitors: progress and potential. Bioorg. Med. Chem. 24, 5017–5027.
LaPorte, M.G., da Paz Lima, D.J., Zhang, F., Sen, M., Grandis, J.R., Camarco, D., Hua, Y.,
Johnston, P.A., Lazo, J.S., Resnick, L.O., Wipf, P., Huryn, D.M., 2014. 2-
Guanidinoquinazolines as new inhibitors of the STAT3 pathway. Bioorg. Med. Chem.
Lett. 24, 5081–5085.
Leach, A.R., Hann, M.M., Burrows, J.N., Griffen, E.J., 2006. Fragment screening: an in-
troduction. Mol. BioSyst. 2, 429–446.
Dahlin, J.L., Walters, M.A., 2016. How to triage PAINS-full research. Assay Drug Dev.
Leeson, P.D., Springthorpe, B., 2007. The influence of drug-like concepts on decision-
making in medicinal chemistry. Nat. Rev. Drug Discov. 6, 881–890.
Li, Y., Wu, C., Huang, J., Su, W., 2006. Mild and convenient synthesis of 1,2-dihy-
droquinolines from anilines and acetone catalyzed by ytterbium(III) triflate in ionic
liquids. Synth. Commun. 36, 3065–3073.
Li, C., Fan, S., Owonikoko, T.K., Khuri, F.R., Sun, S.Y., Li, R., 2011. Oncogenic role of
EAPII in lung cancer development and its activation of the MAPK–ERK pathway.
Oncogene 30, 3802–3812.
Technol. 14, 168–174.
Do, P.M., Varanasi, L., Fan, S., Li, C., Kubacka, I., Newman, V., Chauhan, K., Daniels, S.R.,
Boccetta, M., Garrett, M.R., Li, R., Martinez, L.A., 2012. Mutant p53 cooperates with
ETS2 to promote etoposide resistance. Genes Dev. 26, 830–845.
Emsley, P., Lohkamp, B., Scott, W.G., Cowtan, K., 2010. Features and development of
Coot. Acta Cryst. D 66, 486–501.
Fortune, J.M., Osheroff, N., 2000. Topoisomerase II as a target for anticancer drugs: when
enzymes stop being nice. Prog. Nucleic Acid Res. Mol. Biol. 64, 221–253.
Li, G., Liu, H., Lv, G., Wang, Y., Fu, Q., Tang, Z., 2015. Enantioselective organocatalytic
78

C.J.A. Ribeiro et al.
EuropeanJournalofPharmaceuticalSciences118(2018)67–79
transfer hydrogenation of 1,2-dihydroquinoline through formation of aza-o-xylylene.
Org. Lett. 17, 4125–4127.
Lipinski, C.A., Lombardo, F., Dominy, B.W., Feeney, P.J., 2001. Experimental and com-
putational approaches to estimate solubility and permeability in drug discovery and
development settings. Adv. Drug Deliv. Rev. 46, 3–26.
Llona-Minguez, S., Höglund, A., Wiita, E., Almlöf, I., Mateus, A., Calderón-Montaño, J.M.,
Cazares-Körner, C., Homan, E., Loseva, O., Baranczewski, P., Jemth, A.-S., Häggblad,
M., Martens, U., Lundgren, B., Artursson, P., Lundbäck, T., Jenmalm Jensen, A.,
Warpman Berglund, U., Scobie, M., Helleday, T., 2017. Identification of triazo-
lothiadiazoles as potent inhibitors of the dCTP pyrophosphatase 1. J. Med. Chem. 60,
2148–2154.
Maciejewski, S., Nguyen, J.H.C., Gomez-Herreros, F., Cortes-Ledesma, F., Caldecott, K.W.,
Semler, B.L., 2016. Divergent requirement for a DNA repair enzyme during en-
terovirus infections. MBio 7.
by-products in 3-component synthesis. J. Heterocycl. Chem. 9 (637-&).
Schellenberg, M.J., Appel, C.D., Adhikari, S., Robertson, P.D., Ramsden, D.A., Williams,
R.S., 2012. Mechanism of repair of 5′-topoisomerase II–DNA adducts by mammalian
tyrosyl-DNA phosphodiesterase 2. Nat. Struct. Mol. Biol. 19, 1363–1371.
Schrödinger, 2014a. Schrödinger Release 2014-3: LigPrep, Version 3.1. Schrödinger, LLC,
New York, NY.
Schrödinger, 2014b. Schrödinger Release 2014-3: Schrödinger Suite 2014-3. (a) Protein
Preparation Wizard. (b) Epik version 2.9; Schrödinger, LLC: New York, NY, 2014. (c)
Impact version 6.4; Schrödinger, LLC: New York, NY, 2014. (d) Prime version 3.7;
Schrödinger, LLC: New York, NY.
Shi, K., Kurahashi, K., Gao, R., Tsutakawa, S.E., Tainer, J.A., Pommier, Y., Aihara, H.,
2012. Structural basis for recognition of 5′-phosphotyrosine adducts by Tdp2. Nat.
Struct. Mol. Biol. 19, 1372–1377.
Shikhaliev, K.S., Kryl'skii, D.V., Shestakov, A.S., Falaleev, A.V., 2003. Condensation of
isatoic anhydride with hetarylguanidines. Russ. J. Gen. Chem. 73, 1147–1150.
Shultz, M.D., 2013. Setting expectations in molecular optimizations: strengths and lim-
itations of commonly used composite parameters. Bioorg. Med. Chem. Lett. 23,
5980–5991.
Maede, Y., Shimizu, H., Fukushima, T., Kogame, T., Nakamura, T., Miki, T., Takeda, S.,
Pommier, Y., Murai, J., 2014. Differential and common DNA repair pathways for
topoisomerase I- and II-targeted drugs in a genetic DT40 repair cell screen panel. Mol.
Cancer Ther. 13, 214–220.
Marchand, C., Abdelmalak, M., Kankanala, J., Huang, S.Y., Kiselev, E., Fesen, K.,
Kurahashi, K., Sasanuma, H., Takeda, S., Aihara, H., Wang, Z., Pommier, Y., 2016.
Deazaflavin inhibitors of tyrosyl-DNA phosphodiesterase 2 (TDP2) specific for the
human enzyme and active against cellular TDP2. ACS Chem. Biol. 11, 1925–1933.
McCoy, A.J., Grosse-Kunstleve, R.W., Adams, P.D., Winn, M.D., Storoni, L.C., Read, R.J.,
2007. Phaser crystallographic software. J. Appl. Crystallogr. 40, 658–674.
Menon, V., Povirk, L.F., 2016. End-processing nucleases and phosphodiesterases: an elite
supporting cast for the non-homologous end joining pathway of DNA double-strand
break repair. DNA Repair (Amst) 43, 57–68.
Simmons, K.J., Chopra, I., Fishwick, C.W., 2010. Structure-based discovery of anti-
bacterial drugs. Nat. Rev. Microbiol. 8, 501–510.
Soubhye, J., Chikh Alard, I., Aldib, I., Prévost, M., Gelbcke, M., De Carvalho, A.,
Furtmüller, P.G., Obinger, C., Flemmig, J., Tadrent, S., Meyer, F., Rousseau, A., Nève,
J., Mathieu, V., Zouaoui Boudjeltia, K., Dufrasne, F., Van Antwerpen, P., 2017.
Discovery of novel potent reversible and irreversible myeloperoxidase inhibitors
using virtual screening procedure. J. Med. Chem. 60, 6563–6586.
Theiling, L.F., McKee, R.L., 1952. 2-Guanidinoquinazolines. J. Am. Chem. Soc. 74,
1834–1836.
Murshudov, G.N., Vagin, A.A., Dodson, E.J., 1997. Refinement of macromolecular
structures by the maximum-likelihood method. Acta Crystallogr. D Biol. Crystallogr.
53, 240–255.
Thomson, G., Watson, A., Caldecott, K., Denneny, O., Depledge, P., Hamilton, N.,
Hopkins, G., Jordan, A., Morrow, C., Raoof, A., Waddell, I., Ogilvie, D., 2013.
Generation of assays and antibodies to facilitate the study of human 5′-tyrosyl DNA
phosphodiesterase. Anal. Biochem. 436, 145–150.
Nitiss, J.L., 2009. DNA topoisomerase II and its growing repertoire of biological functions.
Nat. Rev. Cancer 9, 327–337.
Otwinowski, Z., Minor, W., 1997. Processing of X-ray diffraction data collected in oscil-
lation mode. Methods Enzymol. 276, 307–326.
Umamaheswari, A., Pradhan, D., Hemanthkumar, M., 2010. Identification of potential
Leptospira phosphoheptose isomerase inhibitors through virtual high-throughput
screening. Genomics Proteomics Bioinforma. 8, 246–255.
Pan, P., Tian, S., Sun, H., Kong, X., Zhou, W., Li, D., Li, Y., Hou, T., 2015. Identification
and preliminary SAR analysis of novel type-I inhibitors of TIE-2 via structure-based
virtual screening and biological evaluation in in vitro models. J. Chem. Inf. Model.
55, 2693–2704.
Pei, H., Yordy, J.S., Leng, Q., Zhao, Q., Watson, D.K., Li, R., 2003. EAPII interacts with
ETS1 and modulates its transcriptional function. Oncogene 22, 2699–2709.
Pinelli, A., Trivulzio, S., Pojaga, G., Rossoni, G., 1996. Effects of 2-guanidine-4-methyl-
quinazoline on gastric acid secretion in rats. Pharmacol. Res. 34, 225–230.
Pommier, Y., 2013. Drugging topoisomerases: lessons and challenges. ACS Chem. Biol. 8,
82–95.
Pommier, Y., Huang, S.-Y.N., Gao, R., Das, B.B., Murai, J., Marchand, C., 2014. Tyrosyl-
DNA-phosphodiesterases (TDP1 and TDP2). DNA Repair 19, 114–129.
Pommier, Y., Sun, Y., Huang, S.N., Nitiss, J.L., 2016. Roles of eukaryotic topoisomerases
in transcription, replication and genomic stability. Nat. Rev. Mol. Cell Biol. 17,
703–721.
Virgen-Slane, R., Rozovics, J.M., Fitzgerald, K.D., Ngo, T., Chou, W., van Noort, G.J.V.,
Filippov, D.V., Gershon, P.D., Semler, B.L., 2012. An RNA virus hijacks an incognito
function of a DNA repair enzyme. Proc. Natl. Acad. Sci. U. S. A. 109, 14634–14639.
Wang, P., Elsayed, M.S.A., Plescia, C.B., Ravji, A., Redon, C.E., Kiselev, E., Marchand, C.,
Zeleznik, O., Agama, K., Pommier, Y., Cushman, M., 2017. Synthesis and biological
evaluation of the first triple inhibitors of human topoisomerase 1, tyrosyl–DNA
phosphodiesterase 1 (tdp1), and tyrosyl–DNA phosphodiesterase 2 (tdp2). J. Med.
Chem. 60, 3275–3288.
Webb, T.R., Lvovskiy, D., Kim, S.-A., Ji, X.-D., Melman, N., Linden, J., Jacobson, K.A.,
2003. Quinazolines as adenosine receptor antagonists: SAR and selectivity for A2B
receptors. Bioorg. Med. Chem. 11, 77–85.
Wu, Y., Hu, J., Ding, H., Chen, L., Zhang, Y., Liu, R., Xu, P., Du, D., Lu, W., Liu, J., Liu, Y.,
Liu, Y.-C., Lu, J., Zhang, J., Yao, Z., Luo, C., 2016. Identification of novel EZH2 in-
hibitors through pharmacophore-based virtual screening and biological assays.
Bioorg. Med. Chem. Lett. 26, 3813–3817.
Pype, S., Declercq, W., Ibrahimi, A., Michiels, C., Van Rietschoten, J.G., Dewulf, N., de
Boer, M., Vandenabeele, P., Huylebroeck, D., Remacle, J.E., 2000. TTRAP, a novel
protein that associates with CD40, tumor necrosis factor (TNF) receptor-75 and TNF
receptor-associated factors (TRAFs), and that inhibits nuclear factor-kappa B acti-
vation. J. Biol. Chem. 275, 18586–18593.
Rao, T., Gao, R., Takada, S., Al Abo, M., Chen, X., Walters, K.J., Pommier, Y., Aihara, H.,
2016. Novel TDP2-ubiquitin interactions and their importance for the repair of to-
poisomerase II-mediated DNA damage. Nucleic Acids Res. 44, 10201–10215.
Raoof, A., Depledge, P., Hamilton, N.M., Hamilton, N.S., Hitchin, J.R., Hopkins, G.V.,
Jordan, A.M., Maguire, L.A., McGonagle, A.E., Mould, D.P., Rushbrooke, M., Small,
H.F., Smith, K.M., Thomson, G.J., Turlais, F., Waddell, I.D., Waszkowycz, B., Watson,
A.J., Ogilvie, D.J., 2013. Toxoflavins and deazaflavins as the first reported selective
small molecule inhibitors of tyrosyl-DNA phosphodiesterase II. J. Med. Chem. 56,
6352–6370.
Xiao, X., Zhu, M.X., Xu, T.-L., 2013. 2-Guanidine-4-methylquinazoline acts as a novel
competitive antagonist of A type γ-aminobutyric acid receptors. Neuropharmacology
75, 126–137.
Xu, Q., Sun, M., Bai, Z., Wang, Y., Wu, Y., Tian, H., Zuo, D., Guan, Q., Bao, K., Wu, Y.,
Zhang, W., 2017. Design, synthesis and bioevaluation of antitubulin agents carrying
diaryl-5,5-fused-heterocycle scaffold. Eur. J. Med. Chem. 139, 242–249.
Yu, Y., Chen, Z., Li, W.-G., Cao, H., Feng, E.-G., Yu, F., Liu, H., Jiang, H., Xu, T.-L., 2010. A
nonproton ligand sensor in the acid-sensing ion channel. Neuron 68, 61–72.
Zeng, Z., Cortés-Ledesma, F., El Khamisy, S.F., Caldecott, K.W., 2011. TDP2/TTRAP is the
major 5′-tyrosyl DNA phosphodiesterase activity in vertebrate cells and is critical for
cellular resistance to topoisomerase II-induced DNA damage. J. Biol. Chem. 286,
403–409.
Zhang, J.H., Chung, T.D., Oldenburg, K.R., 1999. A simple statistical parameter for use in
evaluation and validation of high throughput screening assays. J. Biomol. Screen. 4,
67–73.
Zhang, L.-X., Zhang, A.-J., Chen, X.-X., Lei, X.-X., Nan, X.-Y., Chen, D.-Y., Zhang, Z.-Y.,
2002. Synthesis and biological activity of 3-(2-furanyl)-6-aryl-1,2,4-triazolo[3,4-b]-
1,3,4–thiadiazoles. Molecules 7, 681.
Ravindranathan, K.P., Mandiyan, V., Ekkati, A.R., Bae, J.H., Schlessinger, J., Jorgensen,
W.L., 2010. Discovery of novel fibroblast growth factor receptor 1 kinase inhibitors
by structure-based virtual screening. J. Med. Chem. 53, 1662–1672.
Rosowsky, A., Modest, E.J., 1972. Chemical and biological studies on dihydro-s-triazines
.16. NMR evidence for formation of 2-guanidino-4-methylquinazolines as anomalous
79