Structure-activity relationships of 2-arylquinazolin-4-ones as highly selective and potent inhibitors of the tankyrases

Article abstract of DOI:10.1016/j.ejmech.2016.04.041

Tankyrases (TNKSs), members of the PARP (Poly(ADP-ribose)polymerases) superfamily of enzymes, have gained interest as therapeutic drug targets, especially as they are involved in the regulation of Wnt signalling. A series of 2-arylquinazolin-4-ones with varying substituents at the 8-position was synthesised. An 8-methyl group (compared to 8-H, 8-OMe, 8-OH), together with a 4′-hydrophobic or electron-withdrawing group, provided the most potency and selectivity towards TNKSs. Co-crystal structures of selected compounds with TNKS-2 revealed that the protein around the 8-position is more hydrophobic in TNKS-2 compared to PARP-1/2, rationalising the selectivity. The NAD⁺-binding site contains a hydrophobic cavity which accommodates the 2-aryl group; in TNKS-2, this has a tunnel to the exterior but the cavity is closed in PARP-1. 8-Methyl-2-(4-trifluoromethylphenyl)quinazolin-4-one was identified as a potent and selective inhibitor of TNKSs and Wnt signalling. This compound and analogues could serve as molecular probes to study proliferative signalling and for development of inhibitors of TNKSs as drugs.

Full text of DOI:10.1016/j.ejmech.2016.04.041

European Journal of Medicinal Chemistry 118 (2016) 316e327
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Research paper
Structure-activity relationships of 2-arylquinazolin-4-ones as highly
selective and potent inhibitors of the tankyrases
Amit Nathubhai a *, Teemu Haikarainen b, Penelope C. Hayward c,
,
d
a
a
b
~
€
Silvia Munoz-Descalzo , Andrew S. Thompson , Matthew D. Lloyd , Lari Lehtio ,
Michael D. Threadgill a
a Medicinal Chemistry, Department of Pharmacy and Pharmacology, University of Bath, Claverton Down, Bath, BA2 7AY, UK
b Biocenter Oulu and Faculty of Biochemistry and Molecular Medicine, University of Oulu, Oulu, Finland
c Department of Genetics, University of Cambridge, Downing Street, Cambridge, CB2 3EH, UK
d Department of Biology and Biochemistry, University of Bath, Claverton Down, Bath, BA2 7AY, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
Tankyrases (TNKSs), members of the PARP (Poly(ADP-ribose)polymerases) superfamily of enzymes, have
gained interest as therapeutic drug targets, especially as they are involved in the regulation of Wnt
signalling. A series of 2-arylquinazolin-4-ones with varying substituents at the 8-position was syn-
thesised. An 8-methyl group (compared to 8-H, 8-OMe, 8-OH), together with a 40-hydrophobic or
electron-withdrawing group, provided the most potency and selectivity towards TNKSs. Co-crystal
structures of selected compounds with TNKS-2 revealed that the protein around the 8-position is
more hydrophobic in TNKS-2 compared to PARP-1/2, rationalising the selectivity. The NADþ-binding site
contains a hydrophobic cavity which accommodates the 2-aryl group; in TNKS-2, this has a tunnel to the
exterior but the cavity is closed in PARP-1. 8-Methyl-2-(4-trifluoromethylphenyl)quinazolin-4-one was
identified as a potent and selective inhibitor of TNKSs and Wnt signalling. This compound and analogues
could serve as molecular probes to study proliferative signalling and for development of inhibitors of
TNKSs as drugs.
Received 13 November 2015
Received in revised form
13 April 2016
Accepted 15 April 2016
Available online 20 April 2016
Keywords:
Tankyrase
PARP
2-Arylquinazolin-4-one
Wnt
Hydrophobic cavity
© 2016 Elsevier Masson SAS. All rights reserved.
1. Introduction
ates telomere repeating binding factor (TRF1) and itself in order to
regulate the elongation of telomeres [4e6]. TNKS-1 is essential for
Tankyrases (TNKSs) are members of the poly(ADP-ribose)
polymerase (PARP) family of enzymes responsible for poly(ADP-
ribosyl)ating acceptor proteins, using NADþ as a substrate [1].
TNKS-1 (PARP-5a, ARTD5) was first reported in 1998, being located
at telomeres [2]. A closely-related isoform, TNKS-2 (PARP-5b,
ARTD6), was first reported in 2001 [3]. TNKS-1 poly(ADP-ribosyl)
the correct structure and function of the mitotic spindle and
centrosome through poly(ADP-ribosyl)ation of nuclear mitotic
apparatus protein (NuMA) [7e9]. TNKS-1 is also involved in the
glucose transport system [10,11]. It has a regulatory role in the life
cycle of viruses, such as EpsteineBarr virus and Herpes simplex
virus [12,13]. Aberrant Wnt signalling is found in the majority of
colon cancers [14]. Tankyrase-1/2 is a component of the Wnt
pathway, where it poly(ADP-ribosyl)ates Axin [14]. Inhibition of
TNKSs causes stabilisation of Axin and degradation of
thus decreasing nuclear -catenin and inhibiting Wnt/
driven proliferation of cancer cells [14]. Increased nuclear accu-
mulation of -catenin is found in many tumours [15e19] and in-
hibition of TNKS and Wnt signalling leads to antitumour activity
in vivo [20,21]. Tankyrases are over-expressed in several clinical
cancers and cancer cell lines, compared with the corresponding
normal tissue/cell type. These tumours include breast cancer [4,22],
colon cancer [23,24], chronic myeloid leukaemia [25], brain
b
-catenin,
Abbreviations: ADP, adenosine diphosphate; Bn, benzyl; Cbz, benzylox-
ycarbonyl; DMAP, 4-dimethylaminopyridine; EDC, N0-ethyl-N-3-(dimethylamino-
propyl)carbodiimide; ES, electrospray; HMBC, heteronuclear multiple-bond
correlation spectroscopy; HSQC, heteronuclear single quantum coherence spec-
troscopy; mESC, mouse embryonic stem cell; MTS, (3-(4,5-dimethylthiazol-2-yl)-5-
(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium); NAD, nicotinamide
adenine dinucleotide; NMR, nuclear magnetic resonance; NuMA, nuclear mitotic
apparatus protein; PARP, poly(ADP-ribose)polymerase; PEG, poly(ethyleneglycol);
THF, tetrahydrofuran; TRF1, telomere repeat binding factor 1; TNKS, tankyrase.
* Corresponding author.
b
b-catenin-
b
E-mail address: a.nathubhai@bath.ac.uk (A. Nathubhai).
http://dx.doi.org/10.1016/j.ejmech.2016.04.041
0223-5234/© 2016 Elsevier Masson SAS. All rights reserved.

A. Nathubhai et al. / European Journal of Medicinal Chemistry 118 (2016) 316e327
317
tumours [26,27], gastric cancers [28] and bladder cancers [29],
pointing to one or more critical roles for the TNKSs in various
tumour types.
measured selectivity. Crystallographic studies rationalise the
binding mode of these molecules, with particular attention to the
nature of the 8-substituent. They reveal that 40-groups which are
hydrophobic and/or electron-withdrawing (Me, Br, Cl, CF3) are
optimal for potent and selective inhibition of the TNKSs compared
to their activity against PARP-1 and PARP-2. To validate the activity
of these novel compounds as inhibitors of potential active onco-
genic Wnt signalling, a mouse embryonic stem cell (mESC) Wnt
reporter assay was used. Our studies reveal that 8-methyl-2-(4-
trifluoromethylphenyl)quinazolin-4-one 18c is a potent and selec-
tive inhibitor of the TNKSs and Wnt signalling.
Two binding sub-sites within the NADþ-binding site have been
identified and targeted for the design and evaluation of TNKS in-
hibitors. One sub-site normally binds the nicotinamide part of
NADþ using the classical PARP motif though H-bonding to Ser1068
and Gly1032 and
p-stacking interactions with Tyr1071 (TNKS-2
numbering). Several of the known inhibitors of the TNKSs bind
here, including flavones 1 [30], 3-arylisoquinolin-1-ones 2 [31] and
aryltetrahydronaphthyridinones 3 (Fig. 1) [32]. In addition, 2-
arylquinazolin-4-one 4, containing a large and hydrophobic 40-
tertbutylphenyl group, binds with the same conformation as 1-3
and is a potent inhibitor of TNKSs [33]. These interactions are also
seen for XAV939 5, an extensively studied early inhibitor of the
TNKSs [34]. IWR-1 6 binds exclusively to the adenosine-binding site
via induction of conformational change of specific amino-acid
residues and does not interact with the nicotinamide-binding site
[35]. IWR-1 6 has been used as a scaffold towards the design and
evaluation of additional inhibitors that bind to the adenosine-
binding site, such as the oxazolidinone derivative 7 [36,37] and
WIKI-4 8 [38]. There is 83% overall sequence homology between
TNKS-1 and TNKS-2 and 94% sequence homology between TNKS-1
and TNKS-2 at the catalytic domain; structures of both domains
have been solved. In many circumstances, these isoforms perform
their biomolecular functions redundantly [39].
2. Results and discussion
2.1. Chemical synthesis
Building on our original preliminary communication [40], we set
out to prepare an extensive library of compounds, using a rational
drug design approach, to assess the correlations between varying
the 40-group carried on the 2-phenyl ring and substitutions at
the ꢀ6, ꢀ7 and ꢀ8 positions of the quinazolin-4-one scaffold with
inhibition of TNKSs. The synthesis of quinazolin-4-ones used clas-
sical methods. 8-Unsubstituted 2-arylquinazolin-4-ones 13a-g,i
were synthesised in moderate yields using one-pot condensation of
anthranilamide 12 with various 4-substituted benzaldehyde de-
rivatives, in the presence of hydrogensulfite ion, with in situ at-
mospheric oxidation of the intermediate aminal (Scheme 1).
Reduction of the aromatic nitro group was achieved by transfer
hydrogenation of 13g to provide the aniline 13h. The 40-methox-
yphenyl group of 13i was demethylated to furnish the corre-
sponding phenol in 13j. In this series, the need for a lactam motif
was tested with the simple amides 14. Acylation of 12 with various
4-substituted benzoyl chlorides gave 14a-g,i. Transfer hydrogena-
tion of 14g gave the aniline contained in 14h but attempted se-
lective demethylation of 14i failed (see Scheme 1).
In a preliminary communication, we previously disclosed a set
of 8-methyl-2-arylquinazolin-4-ones
9
that bind to the
nicotinamide-binding site [40]. The binding modes of these com-
pounds were rationalised using modelling in silico that indicated
specific tolerances around the 2-arylquinazolin-4-one scaffold.
Compounds such as 4, 6 and 9 were used to design dual-binding
inhibitors that are capable of binding to the nicotinamide-binding
site as well as interacting with the adenosine-binding site; these
include 10 and 11 [41,42]. Although our preliminary study provided
some initial information regarding Structure-Activity Relationships
(SARs) of quinazolin-4-ones, to date, a comprehensive study into
the SAR around the 40- and 8-positions of the quinazolin-4-one
scaffold and correlation with inhibition of TNKSs has not been
performed. Here, we report the evaluation of our extended library
of 2-arylquinazolinones with various groups at the 8-position, as
well as a detailed assessment of hydrophobic, hydrophilic, electron-
withdrawing, electron-neutral and electron-donating groups at the
40-position. Counter-screening against PARP-1 and PARP-2
Treatment of 15a,b with 1,10-carbonyldiimidazole at elevated
temperatures formed the intermediates (N-acylimidazoles or isa-
toic anhydrides) which were treated with aqueous ammonia to give
16a,b in good yields. Carbodiimide (EDC) activation and subsequent
treatment with aqueous ammonia also provided 16a,b in excellent
yields [43,44]. The conversion of 16a to 17 and 18 has been briefly
described previously [40]. Acylation of 3-methoxyanthranilamide
16b with 4-substituted benzoyl chlorides provided 19a-g,i. The
40-nitro group of 19g was reduced to the corresponding amine
Fig. 1. Structures of 1-11, reported inhibitors of tankyrases.

318
A. Nathubhai et al. / European Journal of Medicinal Chemistry 118 (2016) 316e327
activity towards the TNKSs. Briefly, formation of the primary am-
O
ides from anthranilic acids 22-24 via EDC activation and treatment
with aqueous ammonia gave 25-27. Compounds 25,26 were acyl-
ated with 4-methylbenzoyl chloride to give the amides 28,29.
Amide 27 was acylated with benzoyl chloride to give 30, after
which 28-30 were cyclised under the usual aqueous basic condi-
tions to provide 31-33 (see Scheme 3).
5
8
4
ii
4a
8a
H
6
7
13g
13i
13h
13j
N
i
3
2'
iii
1'
3'
4'
2
N
1
R4’
CONH2
NH2
13a-j
Compound 39, which contains a CH2NHCbz extension at the 40-
position, was shown in our preliminary study to retain good
inhibitory activity against both TNKSs [40]. The extension is pre-
dicted by studies in silico to extend through the tunnel from the
hydrophobic cavity to the exterior, where the aromatic ring of the
Cbz may interact with the adenosine-binding site. To evaluate
whether the presence of a 8-methyl substituent on the quinazoli-
none makes a significant contribution to binding in the presence of
the CH2NHCbz 40-extension, analogue 38, lacking the 8-Me group,
was prepared along with the related 6,7,8-trimethoxy derivative 40
(Scheme 4). 4-(Aminomethyl)benzoic acid 34 was acylated with
benzyl chloroformate under Schotten-Baumann conditions [45]
before formation of the acid chloride. Subsequent coupling with
anthranilamides 12, 16a, 27 gave 35-37, respectively. Attempted
cyclisation of 35-37 with aqueous sodium hydroxide led to hy-
drolysis of the carbamate; therefore, milder, less nucleophilic, basic
conditions (aq. K2CO3) gave 38-40. The Cbz units were removed by
treatment with HBr, revealing the respective primary amines 41-43
as their hydrobromide salts, to test the tolerability of a polar
charged moiety in the tunnel.
12
iv
O
ii
14g
14h
N
H
R4’
14a-i
a: R4’ = H; b: R4’ = Me; c: R4’ = CF3; d: R4’ = Cl; e: R4’ = Br;
f: R4’ = F; g: R4’ = NO2; h: R4’ = NH2; i: R4’ = OMe; j: R4’ = OH
Scheme 1. Synthesis of 2-arylquinazolin-4-ones 13a-j and acylated anthranilamides
14a-i. Reagents and conditions: i, ArCHO, NaHCO3, AcNMe2, 150 ꢁC, open flask; ii,
þNH4HCOꢀ2 , Pd/C, MeOH, DMF; iii, BBr3, CH2Cl2, reflux, then NaOH; iv, ArCOCl, pyridine,
dry THF.
using transfer hydrogenation conditions as before. Cyclisation of
19a-g,i under strongly basic aqueous conditions (NaOH) gave 20a-
g,i in high yields. Dilute reaction mixtures and elevated tempera-
tures were used to provide efficient cyclisation under basic condi-
tions to the desired products.
Transfer hydrogenation was employed, as before, to reduce the
nitro group of 20g to the aniline in 20h. Demethylation of 20a-g
using BBr3 provided 21a-g in satisfactory yields. Transfer hydro-
genation of the nitro group in 21g gave 21h. Simultaneous deme-
thylation of the 8- and 40-methoxy groups of 20i gave the
corresponding diol 21j (Scheme 2).
Previous studies in silico suggested that substituents may not be
tolerated at the 6- and 7-positions of the quinazolinone core. To test
this experimentally, we prepared 31 and 32, which are
regioisomers of 18b. The 6,7,8-trimethoxyquinazolinone 33 was
also a target in this part of the study to assess whether a trisub-
stituted quinazolin-4-one would retain significant inhibitory
Replacement of
a benzene ring with thiophene is often
considered as a conservative manipulation in drug design. This
replacement was tested in the current series by synthesising and
evaluating the thienopyrimidinone 45. We have previously re-
ported the synthesis of 45 in seven steps from dimethyl 3-
thiahexanedioate in 3.5% overall yield [46]. With the new com-
mercial availability of the aminothiophenecarboxamide 44, we
were able to prepare 45 in one step (Scheme 5). A Cu-catalysed
displacement of the bromine with benzylamine, with oxidative
cyclisation in situ, afforded the target in 16% yield. The ferrocene
derivative 47 was assembled to evaluate the possibility of accom-
modating three-dimensional bulk into the hydrophobic pocket of
TNKSs. Compound 47 was prepared in the usual way (Scheme 5) by
acylation of anthranilamide 12 with ferrocenecarbonyl chloride,
followed by cyclisation of intermediate 46 with under aqueous
basic conditions.
O
H
CO2H
NH2
CONH2
NH2
i
ii
iii
O
N
N
N
H
R
R
Me
Me
R4’
R4’
15a: R = Me
16a: R = Me
17a-i
18a-j
15b: R = OMe
16b: R = OMe
iv
iv
17g
17h
18g
18i
18h
18j
ii
v
O
O
N
H
H
iii
v
N
N
O
N
N
H
OMe
OH
OMe
R4’
R4’
R4’
19a-i
20a-i
21a-j
v
iv
iv
iv
20i
21j
19g
19h
20g
20h
21g
21h
a: R4’ = H; b: R4’ = Me; c: R4’ = CF3; d: R4’ = Cl; e: R4’ = Br; f: R4’ = F; g: R4’ = NO2; h: R4’ = NH2; i: R4’ = OMe; j: R4’ = OH
Scheme 2. Synthesis of 2-arylquinazolin-4-ones 18,20,21 with various 40- and 8-substituents. Reagents and conditions: i, CDI or EDC, aq. NH3, DMF; ii, ArCOCl, pyridine, dry THF; iii,
aq. NaOH (0.5 M), 60 ꢁC; iv, þNH4HCO2ꢀ, Pd/C, MeOH, DMF; v, BBr3, CH2Cl2, reflux, then NaOH.

A. Nathubhai et al. / European Journal of Medicinal Chemistry 118 (2016) 316e327
319
Scheme 5. Synthesis of thienopyrimidinone 45 and ferrocenylquinazolin-4-one 47.
Reagents and conditions: i, K2CO3, CuBr, BnNH2, 125 ꢁC, DMSO, open flask; ii, ferroce-
necarbonyl chloride, pyridine, dry THF; iii, aq. NaOH (0.5 M), 100 ꢁC.
Information). From this initial screen, compounds that dimin-
ished enzyme activity by >60% at 1.0 mM were evaluated further to
Scheme 3. Synthesis of 6- and 7-substituted 2-arylquinazolin-4-ones 31-33. Reagents
and conditions: i, CDI, aq. NH3; ii, ArCOCl, pyridine, dry THF; iii, aq. NaOH (0.5 M).
determine IC50 values against TNKS-1, TNKS-2 and, for counter-
screening, PARP-1. Compounds that showed good selectivity of
inhibition of TNKSs, compared to PARP-1, were also assessed
against PARP-2.
Benzamides 14a-i, 17a-h, 19a-h, 35 and 36 were evaluated
against TNKS-2 at 1.0
mM, 100 nM and 10 nM (Supplementary
Information) to assess the importance of the lactam and its
constraint of the conformation of the H-bond motif, which makes
interactions with Glu1138 and Ser1068 [32]. 1H NMR data suggest
that the primary amide in these compounds is coplanar with the
benzene ring but rotated 180ꢁ relative to its conformation in the
quinazolinones (Fig. 2). Compounds 14a-i, 17a-h, 19a-h, 35 and 36
displayed poor inhibition towards TNKS-2 at 1.0 mM and, therefore,
were not evaluated further. It is thus important that the lactam ring
is fixed to potentiate inhibitory activity towards TNKSs.
The 2-arylquinazolin-4-one derivative 31, bearing a methyl
group at the 7-position, showed weak inhibition against TNKS-2
(18% at 1.0
bearing a 6-methyl group also displayed weak inhibition of TNKS-2
(10% at 1.0 M). Previous studies in silico, accompanied by analysis
mM; see Supplementary Information). Compound 32
m
of the TNKS-2e5 bioactive conformation (PDB code 3KR8), revealed
that substituents in the 7- or 6-positions may be tolerated via po-
tential interaction with Glu1138 but the hydrophobic Me would not
allow this.
In our preliminary communication, we reported that a charged
primary amine was not tolerated at the junction of the hyþdro-
phobic cavity and the tunnel, in that 42 (carrying 40-CH2N H3)
was much less active as an inhibitor of TNKS-2 than was the cor-
responding 40-Me compound 18b [40]. By contrast, more lipophilic
charged tertiary amines are tolerated in this position in the 3-
arylisoquinolin-1-one series [31]. The deleterious effect of the
charged polar 40-CH2NþH3 in the present 2-arylquinazolin-4-one
series was confirmed by the minimal inhibition of TNKS-2 by the
Scheme 4. Synthesis of 2-arylquinazolin-4-ones 38-43 with extensions at the 40-po-
sition. Reagents and conditions: i, CbzCl, aq. K2CO3; ii, SOCl2, iii, 12 or 16a or 27, pyridine,
dry THF; iv, aq. K2CO3 (1.0 M), 100 ꢁC; v, HBr, AcOH.
8-H analogue 41 at 1.0 mM and by the complete inactivity of 43 at
this concentration.
The data for inhibition of the enzymes by the 8-H quinazoli-
nones 13a-j are shown in Table 1. In this series, compounds 13c (40-
CF3), 13d (40-Cl) and 13e (40-Br), which carry lipophilic groups on
the 2-phenyl ring, demonstrate good inhibition of TNKS-1 and
TNKS-2; good isoform-selectivity was noted, with only weak ac-
tivity against PARP-1 and PARP-2. Curiously, the 40-Me analogue
13b showed no isoform-selectivity whatsoever, being a good in-
hibitor of PARP-1 (IC50 ¼ 424 nM). Slightly weaker activity against
the tankyrases was demonstrated by 13h-j, which carry hydrophilic
2.2. Potency and selectivity of inhibition of tankyrases
An initial screen of the inhibitory activity of all test compounds
was performed using three different concentrations of inhibitor
using our previously developed assay, which uses an immobilised
construct of the catalytic and SAM domains of TNKS-2 [40]. Values
are expressed as percentage inhibition of enzyme activity relative
to control containing no inhibitor (see Supplementary

320
A. Nathubhai et al. / European Journal of Medicinal Chemistry 118 (2016) 316e327
Table 2
Inhibition of TNKS-1, TNKS-2, PARP-1 and PARP-2 by 8-OMe 2-arylquinazolin-4-
ones 20. IC50 values in nM.
0
Cpd.
R4
TNKS-1
TNKS-2
PARP-1
PARP-2
Fig. 2. Conformations of aroylanthranilamides 48 and 2-arylquinazolinones 49 in
solution.
20a
20b
20c
20d
20e
20f
20g
20h
20i
H
475 12
153 12
1029 235
167 60
87 34
>10000a
1539 238
Me
CF3
Cl
Br
F
NO2
NH2
OMe
2404 575
>100000a
>10000a
73
155 55
111
2
570 25
6022 1107
259 84
1562 85
928 46
62 11
5
106 23
459 166
863 221
1076 158
128 27
>100000a
>10000a
and/or electron-donating substituents at the 40-position; these
compounds also showed some inhibition of PARP-1, leading to
lower selectivity than 13c-e. Small substituents (H, F) at the 4'-
position in 13a,f lead to lower activity against the tankyrases and
greater affinity for PARP-1. The planar eM nitro group in 13g gives
weaker activity against TNKS-1 but increased affinity for PARP-2,
giving lower selectivity. In this 8-H quinazolinone series, opti-
mum activity and selectivity is seen for non-polar units at the 4'-
position. However, the extended eCH2NHCbz at the 4'-position in
38 provided remarkable selectivity for the TNKS-2 isoform over
TNKS-1, PARP-1 and PARP-2.
299 74
873 334
1284 234
346 25
>10000a
1589 114
1244 29
a
Limited by solubility.
repeated by other workers who demonstrated ca. 11-fold selectivity
for the optimum compound 5-benzamidoisoquinolin-1-one in a
comparative study [49,50].
The presence of an electron-donating/hydrophilic 8-OH group
(21a-i, Table 3) provides more potency towards TNKS, compared to
13a-j (8-H), whereas the IC50 values are broadly similar to those for
20a-i (8-OMe). However, the introduction of this polar group in
13a-j significantly enhances binding to PARP-1, leading to lower
selectivity. As inhibitors 21a-i showed poor selectivity vs. PARP-1,
they were not evaluated further against PARP-2. In this series,
large (Cl, Br) and electron-withdrawing groups (F, CF3, NO2) pro-
vided more potent inhibition of TNKSs, compared to those bearing
small 40-substituents, an observation that mirrors those observed
for compounds presented in Tables 1 and 2 Interestingly, com-
pound 21, presenting an 8-OH substituent, had much greater
antiproliferative activity against HT29 human colon carcinoma cells
(see Supplementary Information), compared to all other quinazo-
lin-4-ones examined. One may speculate whether a requirement
for concurrent inhibition of TNKSs and PARP-1 may be required to
achieve this activity in HT29 colon cancer cells or that the com-
pounds may be taken up into the cells more effectively. However,
the observation that the 8-Me compound 18c is capable of
Changing the 8-substituent from 8-H to 8-OMe (Table 2 20a-i)
provided an overall increase of potency of inhibition of tankyrases,
compared to 13a-j. The nature of the 40-group for compounds 20a-i
shows similar SAR as for 13a-j. Large and electron-withdrawing
groups (20c-g) provide potent and selective inhibition of TNKS.
Again, in this series (Table 2), selective inhibition of TNKS is
observed with small electron-withdrawing 40-groups (40-F in 20f)
then small neutral groups (40-H in 20a), when comparing inhibition
values against those for PARP-1. However, in this series, the pres-
ence of an electron-donating 8-OMe group provides a notable in-
crease in inhibitory potency against PARP-2. For example, 20c,e,g
show sub-micromolar inhibition of PARP-2, giving poor selectivity
of inhibition of the target tankyrases. However, these compounds
may represent a new starting-point for the design of PARP-2-
selective inhibitors, with 20e showing >300-fold selectivity for
PARP-2 over PARP-1. It is notable that previously reported selective
inhibitors of PARP-2 were claimed to have selectivity of only ca. 50-
fold (for 5-benzoyloxyisoquinolin-1-one) [48], which could not be
Table 1
Inhibition of TNKS-1, TNKS-2, PARP-1 and PARP-2 by 8-H 2-arylquinazolin-4-ones 13. IC50 values in nM.
0
Cpd.
R4
TNKS-1
10
1406 156
346 25
425 177
485 50
224 49
781 125
1100 49
786 162
250 70
528 76
2423 1825
TNKS-2
17
2200 32
540 195
133 82
PARP-1
PARP-2
XAV939 5
13a
13b
13c
13d
13e
3
6
1200 23
2863 1753
424 215
>100000a
>100000a
>100000a
>100000a
>10000a
1295 121
2718 1846
6429 769
>10000a
H
Me
CF3
>10000a
Cl
Br
F
54
4
2600 124
>100000a
4688 918
2064 108
101 29
1102 95
259 125
690 160
611 125
189 50
45 24
13f
13g
13h
13i
13j
38
NO2
NH2
OMe
OH
CH2NHCbz
>10000a
a
Limited by solubility.

A. Nathubhai et al. / European Journal of Medicinal Chemistry 118 (2016) 316e327
321
Table 3
importance of the flexibility of the tetrahydrothiapyran in 5. The
ferrocenylquinazolin-4-one 47 displayed weak inhibition of the
Inhibition of TNKS-1, TNKS-2, PARP-1 and PARP-2 by 8-OH 2-arylquinazolin-4-ones
21. IC50 values in nM.
TNKSs (IC50 vs.TNKS-1
2 ¼ 4589 966 nM); weaker binding to PARP-1 was noted
(IC50 10 M) but PARP-2 was inhibited with
¼
1217
28 nM, IC50 vs. TNKS-
>
m
IC50 ¼ 1600 114 nM). Similar activity has been reported previ-
ously by us for 3-ferrocenyl-5-methylisoquinolin-1-one [31]. These
data show that the hydrophobic pocket in the tankyrases is capable
of accepting some 3-D bulk.
Several SAR themes are evident from this comprehensive study.
Firstly, the hydrophobic 8-Me group provides increased potency,
compared to 8-H, 8-OMe and 8-OH, regardless of the nature of the
40-substituent. The SAR at the 4'-substituent for the 8-Me in 18a-j is
comparable to those in this region in 13a-j, 20a-i and 21a-h,j,
where large and electron-withdrawing 40-groups are required for
potent TNKS inhibition. Again, small 40-electron-withdrawing
groups (40-F in 18f) provides similar potency against TNKSs as
compounds bearing small 40-neutral groups (40-H in 18a). However,
as with 13f, 20f and 21f (all 40-F),18f provides selective inhibition of
TNKSs compared to their respective 40-H counterparts. Compound
18e, containing a large 40-Br, is the most potent and selective TNKS
inhibitor in this series. The 4'-extended Cbz group of 39 retains
selective inhibition towards TNKSs over PARP-1/-2.
0
Cpd.
R4
TNKS-1
TNKS-2
PARP-1
21a
21b
21c
21d
21e
21f
21g
21h
21j
H
195 21
180 94
1034 116
559 34
3105 172
645 221
130 40
1361 608
105
289 135
856 137
Me
CF3
Cl
22
2
44
25
6
1
53 11
46 12
49 14
15
Br
F
47
3
1
101 39
58 23
83 18
156 28
263 134
191 52
219 51
86 10
NO2
NH2
OH
5
aLimited by solubility.
inhibiting Wnt signalling in cells (see below) suggests that this
series can be taken up into cells readily.
The inhibitory activities of 18a-j, 39 and 42 (8-Me, Table 4)
against TNKS-1, TNKS-2 and PARP-1 were reported in our pre-
liminary communication [40] and are provided for comparison.
Compounds that did not display PARP-1 inhibitory activity up to
5000 nM have now been further assessed for PARP-2 inhibition to
assess their isoform-selectivity further. Inhibition of PARP-2 by
these compounds was broadly similar to inhibition of PARP-1 but
with somewhat lower IC50 values. The excellent isoform selectivity
for TNKSs vs. other PARPs is further confirmed.
2.3. Structural studies
To rationalise SARs displayed in Tables 1e4, selected compounds
were co-crystallised with the catalytic domain of TNKS-2. The
structures of nine compounds (13b, 13c, 18b, 18c, 20b, 20c, 21b, 21c
and 38) in complex with this construct were solved at 1.7 to 2.4 Å
resolution (Fig. 3), varying the 8-substituent but maintaining the 4'-
group as either 4'-methyl (13b, 18b, 20b, 21b) or 4'-trifluoromethyl
(13c, 18c, 20c, 21c). In all cases, the quinazolin-4-one scaffold binds
into the nicotinamide-binding site forming the usual conserved
interactions. The lactam NeH forms a hydrogen bond to the back-
bone carbonyl of Gly1032, whilst the C]O of the lactam forms two
hydrogen bonds to the backbone C]O of Gly1032 and the hydroxy
The quinazolinone core was changed to thieno[3,4-d]pyrimidin-
4(3H)-one in 45. The 7-position of 45 is unsubstituted, so com-
parison with 13a is apposite. The activities of 45 against the iso-
forms are slightly greater (IC50 vs. TNKS-1 ¼ 466 11 nM, IC50 vs.
TNKS-2 ¼ 2455 123 nM, IC50 vs. PARP-1 ¼ >10000 nM, IC50 vs.
side chain of Ser1068. The quinazolinone/lactam ring creates
p-
PARP-2 ¼ 11500
225 nM), reflecting the conservatism of the
benzene-to-thiophene replacement. However, 45 is much less
potent than XAV939 5 against the tankyrases (Table 1), showing the
stacking interactions with Tyr1071. The hydrophobic cavity, lined
with Tyr1050, Ile1075, Pro1034 and Phe1035, can accommodate large
groups at the 40-position, including Br (13e, 18e, 20e and 21e) and
Cl (13d, 18d, 20d and 21d). Compound 38, which contains the
extended 40-CH2NHCO2CH2Ph interacts further with Pro1034 and
Phe1035, unlike the simpler quinazolinones. The binding of 38 does
not cause large changes in the conformation of the protein (Fig. 3
panel I) but there is some minor structural movement of the
side-chain of Ile1075 as the backbone amide NH of Ile1075 makes a
partial hydrogen bond (3.71 Å; not shown for clarity) to the benzyl
oxygen of 38. No additional or significant movement is observed for
residues that are in close proximity to the large 40-CH2NHCbz group
of 38 as it threads treads through the tunnel at the distal end of the
hydrophobic cavity.
Table 4
Inhibition of TNKS-1, TNKS-2, PARP-1 and PARP-2 by 8-OH 2-arylquinazolin-4-ones
18. IC50 values in nM.
0
Cpd. R4
TNKS-1
TNKS-2
PARP-1
PARP-2
Selective inhibition of TNKSs over PARP-1 can be rationalised by
structural alignment of the TNKS-2-38 structure with a PARP-1
structure containing 8-hydroxy-2-methylquinazolin-4-one (PDB
code 4PAX) (Fig. 3, panel J). The two protein structures were aligned
to overlap the nicotinamide-binding motifs and the structure of the
ligand 8-hydroxy-2-methylquinazolin-4-one has been removed for
clarity in the figures. Fig. 3, panel J shows that the 2-aryl unit of 38
would cause steric clashes with a-helical residues Asn767, Asp766
and Gln763 of PARP-1. Additionally, the cavity accommodating the
40-carboxybenzyl group is in a closed state in PARP-1, whereas the
corresponding region of protein in TNKS-2 is hydrophobic and open
to solvent through the tunnel (Fig. 3, panel I). The hydrophilic na-
ture of this PARP-1 surface is, therefore, suited towards 2-
18a
18b
18c
18d
18e
18f
18g
18h
18i
18j
39
H
Me
CF3
Cl
Br
F
NO2
NH2
41 18a 42 18a 795 48a
35 12a
48 ꢁ7a
36 4a
32 4a
7
2a
685 89a
>5000a,b
>5000a,b
>5000a,b
>5000a,b
10 3a
33 7a
3098 1524
6350 704
10529 3322
683 15
6
1a
44 13a 40 8a
32 8a
35 10a
37 7a
52 11a
44 4a
30 14a 104 46a
3
1a
12 4a
1a
504 128a
891 142a
1200 160a
OMe
OH
6
CH2NHCbz
73 28a >5000a,b
>100000b
42
CH2NþH3 Brꢀ 907 26
43 13
266 19
a
Determined previously [40].
Limited by solubility.
b

322
A. Nathubhai et al. / European Journal of Medicinal Chemistry 118 (2016) 316e327
Fig. 3. Crystal structures of Panel A 13b, Panel B 13c, Panel C 20b, Panel D 20c, Panel E 21b, Panel F 21c, Panel G 18b, Panel H 18c, Panel I 38 bound to TNKS-2. Overlay of 38 onto
PARP-1 (panel J). Key residues are in purple and key H-bond interactions are shown as dashed lines. (For interpretation of the references to colour in this figure legend, the reader is
referred to the web version of this article.)
arylquinazolin-4-one derivatives containing hydrophilic 4'-sub-
stituents (40-NH2 and 40-OH). Conversely, this hydrophilic
-helical
improved inhibitory activity and selectivity towards TNKSs.
The TNKS-2 structure was solved with two molecules in the unit
cell and, therefore, the numbers of water molecules in close prox-
imity to the quinazolinone 8-position in both monomer chains A
and B were considered. The number of water molecules in this area
was averaged between the two chains. We measured the number
and distance of water molecules that were within 6.5 Å of the 8-
carbon of the quinazolinone scaffold for compounds 4b, 10b, 11b
and 12b. There appears to be correlation between the amount and
distance of water molecules and inhibitory activity of these com-
pounds. This is particularly seen between compounds 4b and 10b.
The presence of a total of six water molecules (three in each chain)
for 4b results in a ten-fold decrease in inhibitory activity, compared
with 10b which contains four water molecules (two in each chain)
that are in proximity to C-8 of the compound. A combination of the
hydrophobic nature of protein surrounding the 8-position in TNKS-
1 and TNKS-2 (compared to PARP-1 which is more hydrophilic) and
the amount of water repulsion in this region may provide an
explanation why 8-Me is optimum for potency against TNKSs.
Most of the quinazolinones reported here were more potent
towards PARP-2 than against PARP-1. XAV939 5, which also binds to
the nicotinamide-binding site, was eight-fold more potent towards
PARP-2 compared to PARP-1, which was consistent with literature
values [33].
a
region in PARP-1 is intolerant of electronegative or large hydro-
phobic groups at the 4'-position, leading to greater selectivity of 2-
arylquinazolinones carrying 40-hydrophobic groups towards TNKS-
1 and TNKS-2. The presence of a large, hydrophobic 40-group
causing potent TNKS inhibition has been observed using 4 [33]. A
clear SAR at the 8-position was observed for inhibition of both
TNKS-1 and TNKS-2: 8-Me > 8-OH > 8-OMe > 8-H. Similar SAR had
been deduced earlier for the 3-arylisoquinolin-1-ones (5-Me > 5-
OMe > 5-OH ¼ 5-NH2 ¼ 5-F). Structural alignment of the PARP-1
(PDB code 4PAX) and TNKS-2-38 crystal structures shows that a
larger space around the 8-position is present in PARP-1 than in
TNKS-2 (Fig. 3). In addition, the corresponding hydrophobic residue
Tyr1050 of TNKS-2 is absent from PARP-1. Glu988 in PARP-1 is 3 Å
closer to the 8-position than the corresponding Glu1138 of TNKS-2.
The closer proximity of Glu988 and lack of a hydrophobic residue
around the 8-position of the quinazolinone scaffold provides a
more hydrophilic site in PARP-1, compared to TNKS-2, and may
possibly explain why hydrophilic groups in the 8-position (such as
8-OH) is well tolerated in the PARP-1 binding site. The protein
surrounding the 8-position of the quinazolinone scaffold in TNKS-2
is more hydrophobic in nature due to the presence of a proximal
Tyr1050 and, therefore, a hydrophobic 8-Me group in 18a-j provides

A. Nathubhai et al. / European Journal of Medicinal Chemistry 118 (2016) 316e327
323
2.4. Inhibition of Wnt signalling
Methyl-2-(4-trifluoromethylphenyl)quinazolin-4-one 18c was
shown to be more potent than 5 in this Wnt-inhibition assay, in
that it reduced the proportion of cells with GFP fluorescence >8894
units to 32% (Fig. 4, pale green; Table 5), very close to the basal non-
differentiating level for these cells. These effects are also reflected
in the diminution of the median fluorescence values (Table 5).
Analogous quinazolinones 18d,e, 20c, 39 showed lesser inhibition
of the Wnt response in this assay, while 18f, 20d were effectively
inactive. All the test quinazolinones were evaluated for anti-
proliferative activity against HT29 cells (see below and
Supplementary Information). Inhibitor 18c did not display any ac-
tivity in this MTS cell proliferation assay, consistent with results
shown using mESCs.
Inhibitors 5, 18c, 18d, 18e, 18f, 20d, 20c and 39 displayed se-
lective and potent TNKS inhibition and were investigated further to
assess their cellular activity as inhibitors of the Wnt signalling
cascade, using an established assay based on mouse embryonic
stem cells (mESCs) [51,52]. In mESCs, canonical Wnt signalling is
negligible and only becomes active as the cells differentiate
[51,53,54]. As shown in Fig. 4 (grey) and Table 5, E14Tg2A mESCs,
which lack the green fluorescent protein (GFP) gene, showed
minimal fluorescence. Derived TLg2 cells express a GFP reporter
under the control of Wnt signalling. Under standard growth con-
ditions, TLg2 cells do not differentiate and display basal levels of
green fluorescence (Fig. 4, pink; Table 5). When these cells are
cultured in N2B27 medium for 3 d, they differentiate, there is an
activation of the Wnt reporter and the median GFP fluorescence
increases (from 6405 to 8894 units; Fig. 4, blue; Table 5), as
measured by flow cytometry [51]. Inhibitors of the Wnt response
will lead to a decrease in the percentage of cells with fluorescence
above this value (8894 units). The experiment was repeated in the
presence of test quinazolinones 18c-f, 20c,d, 39 (10 mM) or XAV939
5 (control, 10 mM).
2.5. Antiproliferative activity
Compounds 13a-j, 18a-j, 20a-h, 21a-h, 39, 42, 45 and 47 were
assessed for their antiproliferative/cytotoxic effects against HT29
human colon carcinoma cells using an MTS assay. Three concen-
trations of each compound were used (see Supplementary
Information). Inhibitors 13j and 18j showed inhibition of the pro-
liferation of HT29 cells in a dose-dependent manner. These com-
pounds maintain a 40-OH substituent but differ at the 8-position (8-
H and 8-Me, respectively) and display some PARP-1 inhibitory ac-
tivity (IC50 ¼ 6429 nM and IC50 ¼ 1200 nM, respectively). Inhibitors
21a (40-H), 21b (40-Me), 21d (40-Cl), 21e (40-Br), 21h (40-NH2) and
21j (40-OH) all carry a 8-OH group and are capable of inhibiting the
TNKSs and PARP-1. Their enzyme-inhibitory activities may com-
plement their observed concentration-dependent inhibition of
proliferation of HT29 colon cancer cells. The observed correlation
between dual TNKS-1/-2 and PARP-1 inhibitory activity and inhi-
bition of HT29 cells is also demonstrated by 42. We therefore
speculate whether dual TNKS and PARP-1 inhibition may be
necessary for antiproliferative activity against colon cancer cells.
Alternatively, it may be that the limited solubility of some ana-
logues in the assay medium masked their antiproliferative effect by
not allowing a sufficient concentration to be achieved. In general,
we observed turbid solutions for the 8-Me, 8-OMe and 8-H series of
The positive control, XAV939 5, is known to inhibit Wnt sig-
nalling in a different model [14,51] and this effect was confirmed in
our model by diminution to 37% of the proportion of cells with GFP
fluorescence >8894 units (Fig. 4, dark green; Table 5) (compared
with 50% for no drug and 28% for non-differentiating controls). 8-
compounds at 100 mM (see Supplementary Information). The 8-OH
group of inhibitors 21a (40-H), 21b (40-Me), 21d (40-Cl), 21e (40-Br),
21h (40-NH2) and 21j (40-OH) provides increased aqueous solubility,
thus facilitating the observation of antiproliferative activity at
100 mM. Interestingly, potent and selective inhibitors of TNKSs, for
example 18c and 5, did not inhibit proliferation of HT29 cells but do
inhibit Wnt signalling.
3. Conclusions
In this study, the SAR around the 8- and 4'-positions in series of
2-arylquinazolin-4-ones has been explored for inhibition of the
tankyrases and for activity against the isoforms PARP-1 and PARP-2.
At the 8-position, a small hydrophobic group (methyl) gave the best
potency for inhibition of both tankyrases and selectivity in the
counter-screen against the isoforms PARP-1 and PARP-2.
Conversely, a polar H-bond donor (8-OH) at this position caused
some loss of potency against the tankyrases but considerably
increased activity against PARP-1. These data were rationalised by
modelling studies and by comparison of the crystal structures of
complexes with TNKS-2 and the structure of PARP-1. It was evident
that the protein surface in the region of space which accommodates
the 8-substituent is much more polar in PARP-1 than it is in the
tankyrases. This is the first detailed comparison of the binding sites
in these related NADþ-binding enzymes. The conclusions are
consistent with our previous study of 3-arylisoquinolin-1-ones,
where 5-NH2 (equivalent to the 8-position in the quinazolinones)
Fig. 4. Flow-cytometric analysis of TLg2 cells (and E14Tg2A control cells) grown under
the indicated culture conditions. The black line indicates the median GFP fluorescence
intensity of TLg2 cells cultured for 3 d in the presence of N2B27 medium and % is the
percentage of cells found with fluorescence above that level under the various
conditions.

324
A. Nathubhai et al. / European Journal of Medicinal Chemistry 118 (2016) 316e327
Table 5
Summary of the results obtained after flow cytometry analysis in TLg2 cells cultured in the indicated culture conditions. E14Tg2A cells are a non-fluorescent control cell line.
Cell line
Culture condition
Median GFP fluorescence
% Cells above the median GFP fluorescent intensity in cells under differentiation (N2B27)
E14Tg2A
TLg2
TLg2
TLg2
TLg2
0-3SL
1209
6405
8894
7424
6650
0
28
50
37
32
0-3SL (standard)
N2B27 (differentiation)
N2B27 þ XAV939 5
N2B27 þ 18c
led to increased inhibition of PARP-1 and decreased inhibition of
the tankyrases, in comparison with the 5-Me analogues. Substitu-
tion at the 6- and 7-positions of the quinazolinones is not tolerated.
We had previously established that the optimum location for a
substituent on the exocyclic phenyl ring is 4'- [31]. Examination of
the surfaces of the binding pockets in this region for TNKS-2 and
PARP-1 indicated that larger and more hydrophobic substituents
would be accepted in the tankyrases and this proved to be the case
experimentally. Selectivity for inhibition of tankyrases is seen for
40-Cl and 40-Br, hydrophobic substituents which interact well with
the hydrophobic surface in tankyrases but not with the tighter
hydrophilic surface in PARP-1. The 40-CH2NHCbz unit extends into a
tunnel from the cavity towards solvent; this tunnel is lined by both
hydrophobic and hydrophilic residues. Thus optimum potency and
selectivity for inhibition of tankyrases by 2-arylquinazolin-4-ones
is achieved with 8-methyl and 40-hydrophobic groups.
8-Methyl-2-(4-trifluoromethylphenyl)quinazolin-4-one 18c is
identified by this comprehensive study as a highly potent and se-
lective inhibitor of the tankyrases in vitro and in a functional
cellular assay, with potential for further development as a molec-
ular probe for further studies on the roles of the tankyrases. A panel
of compounds with differing isoform-selectivities is now readily
available as a tool for cellular studies on the interplay of the
members of the PARP superfamily.
distillation over Na/benzophenone. All other solvents were pur-
chased from Fisher Scientific. Analytical TLC was performed using
silica gel 60 F254 pre-coated on Al sheets (0.25 mm thickness).
Column chromatography was performed on silica gel 60
(35e70 mm). Melting points were recorded on a Reichert-Jung
Kofler block apparatus and are uncorrected. 1H and 13C NMR
spectra were recorded using Bruker Advance DPX 500 MHz and
400 MHz (1H) instruments. High resolution mass spectra were
determined using electrospray ionisation (ES) on a Bruker Daltonics
MicroTOF instrument and were calibrated with Naþ HCO2ꢀ. The
brine was saturated. Experiments were conducted at ambient
temperature, unless otherwise stated. Solutions in organic solvents
were dried with anhydrous MgSO4. Solvents were evaporated un-
der reduced pressure. Details are given below for one example of
each reaction type; the syntheses and characterisation of analogues
are reported in the Supplementary Information. The 8-
methylquinazolinones 18 were synthesised as reported in our
preliminary communication.40
4.2.2. 2-Phenylquinazolin-4-one (13a)
2-Aminobenzamide 12 (400 mg, 2.9 mmol) was heated with
benzaldehyde (310 mg, 2.9 mmol) and NaHSO3 (454 mg, 4.3 mmol)
at 150 ꢁC for 3.5 h in N,N-dimethylacetamide (3.5 mL) in an open
flask. The cooled mixture was poured into water and the precipitate
was collected by filtration. Chromatography (CH2Cl2 / CH2Cl2/
MeOH 97:3) and recrystallisation (EtOAc) gave 13a (350 mg, 54%) as
white crystals: mp 235e236 ꢁC (lit. [58] mp 235e237 ꢁC); 1H NMR
4. Experimental
((CD3)2SO)
d
7.48e7.59 (4 H, m, Ph 2,4,6-H3 þ 6-H), 7.72 (1 H, d,
4.1. Protein crystallography
J ¼ 8.0 Hz, 8-H), 7.81 (1 H, t, J ¼ 8.0 Hz, 7-H), 8.13e8.18 (3 H, m, Ph
3,5-H2 þ 5-H), 12.49 (1 H, s, NH); 13C NMR ((CD3)2SO) (HSQC/
Crystal structures of TNKS-2 were used to study the binding
modes of the compounds. The catalytic domain of TNKS-2 was
expressed, purified and crystallised as previously reported [39].
Inhibitors were soaked for 48 h into the crystals in a well solution
(0.2 M LiSO4, 0.1 M TriseHCl pH 8.5, 24e26% PEG 3350) supple-
HMBC)
d 120.97 (C-4a), 125.83 (5-C), 126.54 (6-C), 127.44 (Ph 3,5-
C2), 128.56 (Ph 2,6-C2), 131.35 (Ph 4-C), 132.70 (Ph 1-C), 134.54 (7-
C), 148.69 (8a-C), 152.27 (2-C), 162.17 (4-C).
mented with the compound (100 mM) and 250 mM NaCl. Before
4.2.3. 2-(4-Trifluoromethylphenyl)quinazolin-4-one (13c)
collection of the data, the crystals were briefly soaked in a well
solution supplemented with 20% glycerol and flash frozen in liquid
nitrogen. Data was collected at the Diamond Light Source (Didcot,
UK) at beamlines I04-1 and I03 and at ESRF (Grenoble, France) at
beamline ID23-2. Diffraction data were processed and scaled with
the XDS package [55]. The structures were solved using the Dif-
ference Fourier method with the starting phases derived from the
TNKS-2 structure (PDB accession code 3U9H). REFMAC5 [56] was
used for refinement and COOT [57] for manual building of the
model. Data collection and refinement statistics are shown in the
Supporting Information. Coordinates and structure factors have
been deposited into the RCSB Protein Data Bank (http://www.rcsb.
org) with the following accession codes: 4UHG, 4UI3, 4UI5, 4UI6,
4UI7, 4UI8, 4UFY, 4UFU, 4UI4.
Compound 14c (700 mg, 2.3 mmol) was stirred with aq. NaOH
(0.5 M, 15 mL) at 60 ꢁC for 3.5 h. The mixture was acidified by
addition of aq. HCl (9 M) to pH 2. The precipitate was collected by
filtration and dried under vacuum to give 13c (620 mg, 94%) as a
white solid: mp 305e308 ꢁC (lit. [59] mp 306e308 ꢁC); 1H NMR
((CD3)2SO)
d
7.56 (1 H, t, J ¼ 7.5 Hz, 6-H), 7.78 (1 H, d, J ¼ 7.5 Hz, 8-
H), 7.87 (1 H, m, 7-H), 7.30 (2 H, d, J ¼ 8.0, Ph 3,5-H2), 8.18 (1 H, dd,
J ¼ 8.0, 1.0 Hz, 5-H), 8.37 (2 H, d, J ¼ 8.0 Hz, Ph 2,6-H2), 12.75 (1 H, s,
NH); 13C NMR ((CD3)2SO) (HSQC/HMBC)
d 121.21 (4a-C), 123.96 (q,
J ¼ 270.3 Hz, CF3), 125.50 (q, J ¼ 3.5 Hz, Ph 3,5-C2), 125.12 (5-C),
127.12 (6-C), 127.69 (8-C), 128.73 (Ph 2,6-C2), 130.98 (q, J ¼ 32.1 Hz,
Ph 4-C), 134.74 (7-C), 136.61 (Ph 1-C), 148.44 (8a-C), 151.17 (2-C),
162.10 (4-C).
4.2.4. 2-(4-Hydroxyphenyl)quinazolin-4-one (13j)
4.2. Chemical synthesis
Compound 13i (467 mg, 1.9 mmol) was boiled under reflux with
BBr3 in CH2Cl2 (1.0 M, 11 mL) for 3 h. The solvent was evaporated.
The residue was stirred with aq. NaOH (2.5 M, 100 mL) for 3 h. The
mixture was acidified by addition of aq. HCl (9 M) to pH 2. The
mixture was extracted with ethyl acetate (3 ꢂ 20 mL). Drying and
4.2.1. General materials and methods
Chemical reagents were purchased from Sigma, Aldrich, Fluka,
Acros, Lancaster and Novabiochem. Dry THF was obtained by

A. Nathubhai et al. / European Journal of Medicinal Chemistry 118 (2016) 316e327
325
evaporation gave 13j (420 mg, 95%) as a white solid: mp
258e261 ꢁC (lit. [60] mp 261e263 ꢁC); 1H NMR ((CD3)2SO)
6.88
brs, CONHH); 13C NMR ((CD3)2SO) (HSQC/HMBC)
111.14 (1-C), 115.61 (5-C), 116.44 (3-C), 128.77 (6-C), 141.55 (2-C),
d
21.03 (Me),
d
(2 H, d, J ¼ 8.5 Hz, Ph 3,5-H2), 7.46 (1 H, m, 6-H), 7.67 (1 H, d,
J ¼ 8.0 Hz, 8-H), 7.80 (1 H, m, 7-H), 8.08 (2 H, d, J ¼ 8.5 Hz, Ph 2,6-
H2), 8.12 (1 H, dd, J ¼ 8.0,1.5 Hz, 5-H),10.15 (1 H, s, OH),12.30 (1 H, s,
150.31 (4-C), 171.19 (C]O).
4.2.9. 2-(4-(Phenylmethoxycarbonylaminomethyl)phenyl)
benzamide (35)
NH); 13C NMR ((CD3)2SO) (HSQC/HMBC)
d
115.34 (Ph 3,5-C2),120.57
(4a-C), 123.19 (Ph 1-C), 125.79 (5-C), 125.91 (6-C), 127.18 (8-C),
129.56 (Ph 2,6-C2), 134.50 (7-C), 149.04 (8a-C), 152.11 (2-C), 160.53
(Ph 4-C), 162.30 (4-C).
Compound 34 (4.00 g, 26.5 mmol) in aq. K2CO3 (1.0 M, 144 mL)
was stirred vigorously with BnOCOCl (4.52 g, 26.5 mmol) for 16 h.
The precipitate was collected by filtration and dried. This material
was stirred with SOCl2 (10 mL) under Ar for 16 h. The evaporation
residue was suspended in CH2Cl2 and filtered. Evaporation gave the
crude acyl chloride, which was used immediately. This acyl chloride
in dry THF (70 mL) was stirred with 2-aminobenzamide 12 (1.96 g,
14.4 mmol), pyridine (1.48 g, 18.7 mmol) and DMAP (0.35 g,
2.88 mmol) under Ar for 16 h. The evaporation residue, in EtOAc,
was washed twice with water and twice with brine. Drying, evap-
oration and chromatography (EtOAc/CH2Cl2 2:3) gave 35 (5.1 g,
4.2.5. 2-Benzamidobenzamide (14a)
Dry pyridine (151 mg, 1.2 mmol) was added to 12 (200 mg,
1.47 mmol) in dry THF (4.0 mL), followed by benzoyl chloride
(228 mg, 1.6 mmol) in dry THF (4.0 mL). The mixture was stirred
under Ar for 16 h. Evaporation and chromatography (EtOAc/CH2Cl2
2:3) gave 14a (351 mg, 99%) as a white solid: mp 233e235 ꢁC (lit.
[61] mp 234 ꢁC); 1H NMR ((CD3)2SO)
d 7.17e7.19 (1H, m, 5-H),
7.56e7.59 (3 H, m, 4-H þ Ar 3,5-H2), 7.63 (1 H, m, Ar 4-H), 7.65 (1 H,
br, s, NHH), 7.89 (1 H, dd, J ¼ 8.0, 1.5 Hz, 6-H), 7.94e7.96 (2 H, m, Ar
2,6-H2), 8.43 (1 H, brs, NHH), 8.72 (1 H, dd, J ¼ 8.5, 1.0 Hz, 3-H),
88%) as a white solid: mp 180e183 ꢁC; 1H NMR ((CD3)2SO)
d 4.29
(2 H, d, J ¼ 6.0 Hz, CH2), 5.06 (2 H, s, OCH2), 7.17e7.19 (1 H, m, 4-H),
7.30e7.38 (5 H, m, Cbz Ph 3,5-H2 þ Ph 2,6-H2 þ Ph 4-CH2NH), 7.44
(2 H, d, J ¼ 8.0 Hz, Cbz Ph 2,6-H2), 7.57e7.59 (1 H, m, 5-H), 7.86e7.90
(4 H, m, Ph 2,6-H2 þ CONHH þ H-3), 7.94 (1 H, t, J ¼ 7.5 Hz, Cbz Ph 4-
H), 8.44 (1 H, s, CONHH), 8.70 (1 H, d, J ¼ 8.5 Hz, 6-H), 12.95 (1 H, s,
12.96 (1 H, s, NH); 13C ((CD3)2SO) (HSQC/HMBC)
d 119.15 (1-C),
120.00 (3-C), 122.63 (5-C), 126.92 (Ar 2,6-C2), 128.73 (6-C), 128.92
(Ar 3,5-C2), 132.03 (Ar 4-C), 132.58 (4-C), 140.08 (2-C), 134.60 (Ar 1-
C), 164.37 (NHCO), 171.15 (CONH2).
2-NH); 13C NMR ((CD3)2SO) (HMBC/HSQC)
d 43.58 (CH2NH), 65.55
(OCH2), 119.08 (1-C), 120.02 (6-C), 122.64 (4-C), 127.08 (Ph 3,5-C2),
127.42 (Cbz Ph 2,6-C2), 127.85 (Ph 2,6-C2), 127.89 (3-C), 128.42 (Cbz
3,5-C2), 128.79 (Cbz Ph 4-C), 132.66 (5-C), 133.24 (Ph 1-C), 137.13
(Cbz Ph 1-C), 140.18 (2-C), 144.08 (Ph 4-C), 156.47 (Cbz NHCO),
164.26 (NHCO), 171.20 (CONH2); MS (ES) m/z 426.1437 [M þ Na]þ
(C23H21N3NaO4 requires 426.1430).
4.2.6. 2-Amino-3-methoxybenzamide (16b)
Compound 15b (3.00 g, 17.9 mmol) in dry DMF (80 mL) was
treated with 1,1-carbonyldiimidazole (3.19 g, 19.7 mmol) at 70 ꢁC
under Ar for 1 h, after which aq. NH3 (35%, 50 mL) was added
dropwise and the mixture was stirred for 16 h. The mixture was
cooled to 20 ꢁC and was diluted with EtOAc (100 mL). The mixture
was washed twice with water and twice with brine. Drying and
evaporation gave 16b (2.38 g, 80%) as a white solid: mp 139e141 ꢁC
4.2.10. 2-(4-(Phenylmethoxycarbonylaminomethyl)phenyl)
quinazolin-4-one (38)
(lit. [40] mp 139e141 ꢁC); 1H NMR ((CD3)2SO)
6.23 (2 H, br, AreNH2), 6.44 (1 H, t, J ¼ 8.0 Hz, 5-H), 7.04 (1 H, dd,
J ¼ 7.6, 0.8 Hz, 6-H), 7.03 (1 H, brs, CONHH), 7.16 (1 H, dd, J ¼ 8.0,
1.2 Hz, 4-H), 7.67 (1 H, brs, CONHH); 13C NMR ((CD3)2SO) (HSQC/
d
3.77 (3 H, s, Me),
Compound 35 (100 mg, 0.25 mmol) was stirred with aq. K2CO3
(1.0 M, 50 mL) at 100 ꢁC for 16 h. The cooled reaction mixture was
acidified to pH~1. The precipitate was collected by filtration.
Chromatography (EtOAc/petroleum ether 3:2) gave 38 (90 mg, 94%)
HMBC)
d
55.53 (Me), 111.98 (6-C), 113.38 (3-C), 113.64 (5-C), 120.42
as a white solid: mp 244e248 ꢁC; 1H NMR ((CD3)2SO)
d 4.30 (2 H, d,
(4-C), 140.19 (1-C), 146.88 (2-C), 171.19 (C]O).
J ¼ 6.5 Hz, Ar 4-CCH2), 5.07 (2 H, s, Cbz CH2), 7.32e7.38 (5 H, m, Cbz
Ph-H5), 7.42 (2 H, d, J ¼ 8.5 Hz, Ar 3,5-H2), 7.52 (1 H, m, 6-H), 7.74
(1 H, d, J ¼ 8.0 Hz, 8-H), 7.84 (1 H, m, 7-H), 7.94 (1 H, t, J ¼ 6.0 Hz,
CH2NH), 8.13e8.16 (3H, m, Ar 2,6-H2 þ 5-H), 12.55 (NH); 13C
4.2.7. 2-(4-Aminophenyl)-8-methoxyquinazolin-4-one (20h)
Compound 20g (100 mg, 0.34 mmol) was stirred with Pd/C (10%,
100 mg) and þNH4 HCOꢀ2 (212 mg, 3.4 mmol) in MeOH (6 mL) and
DMF (6 mL) under Ar for 3 h. The mixture was filtered (Celite).
Evaporation and chromatography (EtOAc/petroleum ether 4:1)
gave 20h (65 mg, 83%) as a white solid: mp 271e274 ꢁC (lit. [47] mp
((CD3)2SO) (HMBC/HSQC)
d 44.25 (Ar 4-CCH2), 66.17 (Cbz CH2),
121.62 (4a-C), 126.54 (5-C), 127.19 (6-C), 127.76 (Ar 3,5-C2), 128.13
(8-C), 128.47 (Ar 2,6-C2 þ Cbz 2,6-C2), 128.52 (Cbz 4-C), 129.06 (Cbz
3,5-H2), 131.96 (Ar 1-C), 135.28 (7-C), 137.80 (Cbz 1-C), 144.05 (Ar 4-
C), 149.52 (8a-C), 152.84 (2-C), 157.13 (Cbz-CO), 163.10 (4-C); MS
(ES) m/z 408.1317 [M þ Na]þ (C23H19N3NaO3 requires 408.1324].
263e265 ꢁC); 1H NMR ((CD3)2SO)
d 3.92 (3 H, s, Me), 5.80 (2 H, s,
NH2), 6.63 (2 H, dt, J ¼ 8.5, 2.5 Hz, Ar 3,5-H2), 7.29e7.34 (2 H, m, 6-
H þ 7-H), 7.64 (1 H, dd, J ¼ 7.5, 2.0 Hz, 5-H), 7.95 (2 H, dt, J ¼ 9.5,
2.5 Hz, Ar 2,6-H2), 12.09 (1 H, s, NH); 13C NMR ((CD3)2SO) (HSQC/
4.2.11. 2-(4-(Aminomethylphenyl)quinazolin-4-one hydrobromide
(41)
HMBC)
d 55.97 (OMe), 113.01 (Ar 3,5-C2), 114.91 (7-C), 116.91 (5-C),
119.11 (Ar 1-C), 121.28 (4a-C), 125.38 (6-C), 129.07 (Ar 2,6-C2),
140.05 (8a-C), 151.11 (Ar 4-C), 151.98 (2-C), 154.32 (8-C), 162.30 (4-
C).
Compound 38 (0.15 g, 0.39 mmol) was stirred with HBr in AcOH
(33%, 1.5 mL) for 16 h. Evaporation gave 41 (0.13 g, 92%) as a white
solid: mp 277e285 ꢁC; 1H NMR (CD3OD)
d 4.34 (2 H, s, CH2), 7.76
(1 H, m, 6-H), 7.85 (2 H, d, J ¼ 8.0 Hz, Ar 3,5-H2), 7.90 (1 H, d,
J ¼ 8.5 Hz, 8-H), 8.03 (1 H, m, 7-H), 8.14 (2 H, d, J ¼ 8.0 Hz, Ar 2,6-
H2), 8.36 (1 H, dd, J ¼ 8.0, 1.0 Hz, 5-H); 13C NMR (CD3OD) (HMBC/
4.2.8. 2-Amino-4-methylbenzamide (25)
Compound 22 (3.00 g, 19.8 mmol) in dry DMF (100 mL) under Ar
was treated with EDC.HCl (4.17 g, 21.8 mmol) and HOBt (3.33 g,
21.8 mmol) for 3 h, after which aq. NH3 (35%, 20 mL) was added and
the mixture was stirred for 16 h. The evaporation residue, in EtOAc,
was washed twice with water and with brine. Drying and evapo-
ration gave 25 (2.04 g, 69%) as an off-white solid: mp 151e153 ꢁC
HSQC)
d 43.72 (CH2), 121.30 (4a-C), 122.36 (8-C), 128.32 (5-C),
129.28 (Ar 1-C), 130.23 (6-C), 130.83 (Ar 2,6-C2), 131.00 (Ar 3,5-C2),
137.54 (7-C), 140.85 (Ar 4-C), 141.58 (8a-C), 158.58 (2-C), 161.40 (4-
C); HRMS (ES) m/z 252.1140 [M þ H]þ (C15H14N3O requires
252.1137].
(lit. [62] mp 148 ꢁC); 1H NMR ((CD3)2SO)
d 2.22 (3 H, s, Me), 6.35
(1 H, dd, J ¼ 8.0, 1.2, 5-H), 6.52 (1 H, d, J ¼ 0.5 Hz, 3-H), 6.57 (2 H, s,
4.2.12. 2-Phenylthieno[3,4-d]pyrimidin-4-one (45)
NH2), 6.95 (1 H, brs, CONHH), 7.48 (1 H, d, J ¼ 8.0 Hz, 6-H), 7.65 (1 H,
Compound 44 (580 mg, 2.8 mmol) was stirred with K2CO3

326
A. Nathubhai et al. / European Journal of Medicinal Chemistry 118 (2016) 316e327
[2] S. Smith, I. Giriat, A. Schmitt, T. de Lange, Tankyrase, a poly(ADP-ribose) po-
lymerase at human telomeres, Science 282 (1998) 1484e1487.
[3] R.J. Lyons, R. Deane, D.K. Lynch, Z.-S.J. Ye, G.M. Sanderson, H.J. Eyre,
G.R. Sutherland, R.J. Daly, Identification of a novel human tankyrase through
its interaction with the adaptor protein Grb14, J. Biol. Chem. 276 (2001)
17172e17180.
[4] S. Gelmini, M. Poggesi, V. Distante, S. Bianchi, L. Simi, M. Luconi, C. Casini
Raggi, L. Cataliotti, M. Pazzaglia, C. Orlando, Tankyrase, a positive regulator of
telomere elongation, is over expressed in human breast cancer, Cancer Lett.
216 (2004) 81e87.
[5] S. Zimmermann, U.M. Martens, Telomeres and telomerase as targets for
cancer therapy, Cell. Mol. Life Sci. 64 (2007) 906e921.
[6] M. Seimiya, The telomeric PARP, tankyrases, as targets for cancer therapy, Br. J.
Cancer 94 (2006) 341e345.
[7] P. Chang, M. Coughlin, T.J. Mitchison, Tankyrase-1 polymerization of
poly(ADP-ribose) is required for spindle structure and function, Nat. Cell Biol.
7 (2005) 1133e1139.
(1.17 g, 8.5 mmol), CuBr (40 mg, 0.28 mmol) and BnNH2 (610 mg,
5.7 mmol) at 125 ꢁC in an open flask for 16 h in DMSO (28 mL). The
mixture was diluted with EtOAc (30 mL) and was washed twice
with water and twice with brine. Drying, evaporation, chroma-
tography (CH2Cl2
(CH2Cl2) gave 45 (102 mg, 16%) as a yellow solid: mp 280e283 ꢁC
(lit. [46] 275e277 ꢁC); 1H NMR ((CD3)2SO)
7.50e7.56 (3 H, m, Ar
/ EtOAc/CH2Cl2 1:5) and recrystallisation
d
3,4,5-H3), 7.86 (1 H, d, J ¼ 3.5 Hz, 7-H), 8.10 (2 H, m, Ph 2,6-H2), 8.51
(1 H, d, J ¼ 3.5 Hz, 5-H), 11.93 (1 H, s, NH); 13C NMR ((CD3)2SO)
(HSQC/HMBC)
d 118.32 (7-C), 125.01 (4a-C), 127.61 (Ph 2,6-C2),
127.89 (5-C), 128.50 (Ph 3,5-C2), 131.04 (Ph 4-C), 132.91 (Ph 1-C),
148.72 (7a-C), 150.86 (2-C), 158.91 (4-C).
[8] J.N. Dynek, S. Smith, Resolution of sister telomere association is required for
progression through mitosis, Science 304 (2004) 97e100.
4.3. Enzyme assays
[9] W. Chang, J.N. Dynek, S. Smith, NuMA is a major acceptor of poly(ADP-ribosyl)
ation by tankyrase 1 in mitosis, Biochem. J. 391 (2005) 177e184.
[10] T.-Y.J. Yeh, J.I. Sbodio, Z.-Y. Tsun, B. Luo, N.-W. Chi, Insulin-stimulated
exocytosis of GLUT4 is enhanced by IRAP and its partner tankyrase, Biochem.
J. 402 (2007) 279e290.
Test compounds were assayed for inhibition of tankyrase-1,
tankyrase-2, PARP-1 and PARP-2 activity as described previously
[31,33,40].
[11] H.-L. Guo, C. Zhang, Q. Liu, Q.X. Li, G. Lian, D. Wu, X. Li, W. Zhang, Y. Shen,
Z. Ye, S.-Y. Lin, S.-C. Lin, The Axin/TNKS complex interacts with KIF3A and is
required for insulin-stimulated GLUT4 translocation, Cell Res. 22 (2012)
1246e1257.
4.4. Mouse embryonic stem cell culture and Wnt signalling reporter
assay
[12] Z. Deng, C. Atanasiu, K. Zhao, R. Marmorstein, J.I. Sbodio, N.-W. Chi,
P.M. Leiberman, Inhibition of Epstein-Barr virus OriP function by tankyrase, a
telomere-associated poly-ADP ribose polymerase that binds and modifies
EBNA1, J. Virol. 79 (2005) 4640e4650.
[13] Z. Li, Y. Yamauchi, M. Kamakura, T. Murayama, F. Goshima, H. Kimura,
Y. Nishiyama, Herpes simplex virus requires poly(ADP-ribose) polymerase
activity for efficient replication and induces extracellular signal-related ki-
nase-dependent phosphorylation and ICP0-dependent nuclear localization of
tankyrase 1, J. Virol. 86 (2012) 492e503.
[14] S.-M.A. Huang, Y.M. Mishina, S. Liu, A. Cheung, F. Stegmeier, G.A. Michaud,
O. Charlat, E. Wiellette, Y. Zhang, S. Wiessner, M. Hild, X. Shi, C.J. Wilson,
C. Mickanin, V. Myer, A. Fazal, R. Tomlinson, F. Serluca, W. Shao, H. Cheng,
M. Shultz, C. Rau, M. Schirle, J. Schlegl, S. Ghidelli, S. Fawell, C. Lu, D. Curtis,
M.W. Kirschner, C. Lengauer, P.M. Finan, J.A. Tallarico, T. Bouwmeester,
J.A. Porter, A. Bauer, F. Cong, Tankyrase inhibition stabilizes axin and antag-
onizes Wnt signalling, Nature 461 (2009) 614e620.
The cell lines used in this study were E14Tg2A (a mouse em-
bryonic stem cell (mESC) line) and TLg2 (derived from E14Tg2A and
harboring a Wnt-responsive element driving eGFP) [51,54]. The
activity of the test compounds was assessed as described previ-
ously [51]. Briefly, TLg2 cells were plated on gelatin at low density
(4 ꢂ 104 cells per well in a 6-well plate) in either serum þ LIF
(standard culture conditions) or in the presence of N2B27 (differ-
entiating conditions) in the absence (untreated) or presence
(treated) of test compound (10 mM). The medium was refreshed
daily and the Wnt activity was assessed by flow cytometry using a
LSR Fortessa cytometer (BD Bioscience). Live single cells were
selected on forward and side scatter area, side scatter width vs. area
and DAPI exclusion (405 laser and 450/50 filter set). Expression of
eGFP was assessed using the 488 laser and 530/30 filter set. Data
were analyzed and presented using Flowjo software (TreeStar). No
test compound was toxic during the 3 d of treatment.
[15] B.D. White, A.J. Chien, D.W. Dawson, Dysregulation of Wnt/b-catenin signaling
in gastrointestinal cancers, Gastroenterology 142 (2012) 219e232.
[16] M. Katoh, H. Kirikoshi, H. Terasaki, K. Shiokawa, WNT2B2 mRNA, up-regulated
in primary gastric cancer, is a positive regulator of the WNTeb-catenineTCF
signaling pathway, Biochem. Biophys. Res. Commun. 289 (2001) 1093e1098.
[17] Y. Miyoshi, K. Iwao, Y. Nagasawa, T. Aihara, Y. Sasaki, S. Imaoka, M. Murata,
T. Shimano, Y. Nakamura, Activation of the b-catenin gene in primary hepa-
tocellular carcinomas by somatic alterations involving exon 3, Cancer Res. 58
(1998) 2524e2527.
Acknowledgements
[18] M. Katoh, Expression and regulation of WNT1 in human cancer: up-regulation
of WNT1 by b-estradiol in MCF-7 cells, Int. J. Oncol. 22 (2003) 209e212.
[19] R. Najdi, R.F. Holcombe, M.L. Waterman, Wnt signaling and colon carcino-
genesis: beyond APC, J. Carcinog. 10 (2011) 5.
The authors thank Worldwide Cancer Research for generous
financial support (grant No.10-0104). Protein crystallographic work
was funded by Biocenter Oulu and Academy of Finland (grant No.
266922). The use of the facilities and expertise of the Biocenter
Oulu core facility, a member of Biocenter Finland and Instrut-FI, is
gratefully acknowledged. Protein crystallographic experiments
were performed on the ID23-2 beamline at the European Syn-
chrotron Radiation Facility (ESRF), Grenoble, France and the Dia-
mond Light Source (Didcot, UK) Beamlines I04-1 and I03. We are
grateful to Local Contracts at ESRF and the Diamond Light Source
for providing assistance in using beamlines. The research leading to
these results has received funding from the European Community's
Seventh Framework Programme (FP7/2007e2013) under Bio-
Struct-X (grant agreement No. 283570).
[20] J. Waaler, O. Machon, L. Tumova, H. Dinh, V. Korinek, S.R. Wilson, J.E. Paulsen,
N.M. Pedersen, T.J. Eide, O. Machonova, D. Gradl, A. Voronkov, J.P. von Kries,
S. Krauss, A novel tankyrase inhibitor decreases canonical Wnt signaling in
colon carcinoma cells and reduces tumor growth in conditional APC mutant
mice, Cancer Res. 72 (2012) 2822e2832.
[21] J. Waaler, O. Machon, J.P. von Kries, S.R. Wilson, E. Lundenes, D. Wedlich,
D. Gradl, J.E. Paulsen, O. Machonova, J.L. Dembinski, H. Dinh, S. Krauss, Novel
synthetic antagonists of canonical Wnt signaling inhibit colorectal cancer cell
growth, Cancer Res. 71 (2011) 197e205.
[22] N. Sidorova, L. Zavalishina, S. Kurchashova, N. Korsakova, V. Nazhimov,
G. Frank, A. Kuimov, Immunohistochemical detection of tankyrase 2 in human
breast tumors and normal renal tissue, Cell Tissue Res. 323 (2006) 137e145.
[23] S. Gelmini, M. Poggesi, P. Pinzani, S.C. Mannurita, F. Cianci, R. Valanzano,
C. Orlando, Distribution of tankyrase-1 mRNA expression in colon cancer and
its prospective correlation with progression stage, Oncol. Rep. 16 (2006)
1261e1266.
[24] Y.V. Shebzukhov, I.N. Lavrik, J. Karbach, S.V. Khlgatian, E.P. Koroleva,
P.V. Belousov, K.N. Kashkin, A. Knuth, E. Jager, N.-W. Chi, D.V. Kuprash,
S.A. Nedospasov, Human tankyrases are aberrantly expressed in colon tumors
and contain multiple epitopes that induce humoral and cellular immune re-
sponses in cancer patients, Cancer Immunol. Immunother. 57 (2008)
871e881.
[25] L.J. Campbell, C. Fidler, H. Eagleton, A. Peniket, R. Kusec, S. Gal, T.J. Littlewood,
J.S. Wainscoat, J. Boultwood, hTERT, the catalytic component of telomerase, is
downregulated in the haematopoietic stem cells of patients with chronic
myeloid leukaemia, Leukemia 20 (2006) 671e679.
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.ejmech.2016.04.041.
References
[26] A. Shervington, R. Patel, C. Lu, N. Cruickshanks, R. Lea, G. Roberts, T. Dawson,
L. Shervington, Telomerase subunits expression variation between biopsy
[1] D.V. Ferraris, Evolution of poly(ADP-ribose) polymerase-1 (PARP-1) inhibitors.
From concept to clinic, J. Med. Chem. 53 (2010) 4561e4584.

A. Nathubhai et al. / European Journal of Medicinal Chemistry 118 (2016) 316e327
327
samples and cell lines derived from malignant glioma, Brain Res. 1134 (2007)
45e52.
[27] B. Tang, J. Wang, J. Fang, B. Jiang, M. Zhang, Y. Wang, Z. Yang, Expression of
[44] A. Nathubhai, R. Patterson, T.J. Woodman, H.E.C. Sharp, M.T.Y. Chui,
H.H.K. Chung, S.W.S. Lau, J. Zheng, M.D. Lloyd, A.S. Thompson, M.D. Threadgill,
N3-Alkylation during formation of quinazolin-4-ones from condensation of
anthranilamides and orthoamides, Org. Biomol. Chem. 9 (2011) 6089e6099.
[45] C. Schotten, Uber die Oxydation des Piperidins, Ber. Deuts. Chem. Ges. 17
(1884) 2544e2547.
[46] A.E. Shinkwin, W.J.D. Whish, M.D. Threadgill, Synthesis of thiophenecarbox-
amides, thieno[3,4-c]pyridin-4(5H)-ones and thieno[3,4-d]pyrimidin-4(3H)-
ones and preliminary evaluation as inhibitors of poly(ADP-ribose)polymerase
(PARP), Bioorg. Med. Chem. 7 (1999) 297e308.
[47] R.J. Griffin, S. Srinivasan, K. Bowman, A.H. Calvert, N.J. Curtin, D.R. Newell,
L.C. Pemberton, B.T. Golding, Resistance-modifying agents. 5. Synthesis and
biological properties of quinazolinone inhibitors of the DNA repair enzyme
poly(ADP-ribose) polymerase (PARP), J. Med. Chem. 41 (1998) 5247e5256.
[48] R. Pellicciari, E. Camaioni, G. Costantino, I. Formentini, P. Sabbatini,
F. Venturoni, G. Eren, D. Bellocchi, A. Chiarugi, F. Moroni, On the way to se-
lective PARP-2 inhibitors. Design, synthesis, and preliminary evaluation of a
series of isoquinolinone derivatives, ChemMedChem 3 (2008) 914e923.
[49] P.T. Sunderland, E.C.Y. Woon, A. Dhami, A. Bergin, M.F. Mahon, P.J. Wood,
L.A. Jones, S.R. Tully, M.D. Lloyd, A.S. Thompson, H. Javaid, N.M.B. Martin,
TNKS1 is correlated with pathologic grade and Wnt/b-catenin pathway in
human astrocytomas, J. Clin. Neurosci. 19 (2012) 139e143.
[28] J. Gao, J. Zhang, Y. Long, Y. Tian, X. Lu, Expression of tankyrase 1 in gastric
cancer and its correlation with telomerase activity, Pathol. Oncol. Res. 17
(2011) 685e690.
[29] S. Gelmini, S. Quattrone, F. Malentacchi, D. Villari, F. Travaglini, G. Giannarini,
A. Della Melina, M. Pazzagli, G. Nicita, C. Selli, C. Orlando, Tankyrase-1 mRNA
expression in bladder cancer and paired urine sediment: preliminary expe-
rience, Clin. Chem. Lab. Med. 45 (2007) 862e866.
€
[30] M. Narwal, T. Haikarainen, A. Fallarero, P.M. Vuorela, L. Lehtio, Screening and
structural analysis of flavones inhibiting tankyrases, J. Med. Chem. 56 (2013)
3507e3517.
[31] H.A. Paine, A. Nathubhai, E.C.Y. Woon, P.T. Sunderland, P.J. Wood, M.F. Mahon,
€
M.D. Lloyd, A.S. Thompson, T. Haikarainen, M. Narwal, L. Lehtio,
M.D. Threadgill, Exploration of the nicotinamide-binding site of the tank-
yrases, identifying 3-arylisoquinolin-1-ones as potent and selective inhibitors
in vitro, Bioorg. Med. Chem. 23 (2015) 5891e5908.
[32] K. Kumpan, A. Nathubhai, C. Zhang, P.J. Wood, M.F. Mahon, M.D. Lloyd,
M.D. Threadgill, 5-Benzamidoisoquinolin-1-ones and 5-(u-carboxyalkyl)iso-
quinolin-1-ones as isoform-selective inhibitors of poly(ADP-ribose) poly-
merase 2 (PARP-2), J. Med. Chem. 54 (2011) 2049e2059.
€
A.S. Thompson, T. Haikarainen, L. Lehtio, M.D. Threadgill, Structure-based
design, synthesis and evaluation in vitro of arylnaphthyridinones, arylpyr-
idopyrimidinones and their tetrahydro derivatives as inhibitors of the tank-
yrases, Bioorg. Med. Chem. 23 (2015) 3013e3032.
[50] P.T. Sunderland, A. Dhami, M.F. Mahon, L.A. Jones, S.R. Tully, M.D. Lloyd,
A.S. Thompson, H. Javaid, N.M.B. Martin, M.D. Threadgill, Synthesis of 4-alkyl-,
4-aryl- and 4-arylamino-5-aminoisoquinolin-1-ones and identification of a
new PARP-2 selective inhibitor, Org. Biomol. Chem. 9 (2011) 881e891.
[33] T. Haikarainen, J. Koivunen, M. Narwal, H. Venkannagari, E. Obaji, P. Joensuu,
T. Pihlajaniemi, L. Lehtiӧ, Para-Substituted 2-phenyl-3,4-dihydroquinazolin-4-
ones as potent and selective tankyrase inhibitors, ChemMedChem 8 (2013)
1978e1985.
~
[51] F. Faunes, P. Hayward, S. Munoz-Descalzo, S.S. Chatterjee, T. Balayo, J. Trott,
A. Christoforou, A. Ferrer-Vaquer, A.-K. Hadjantonakis, R. Dasgupta, A.M. Arias,
[34] M. Narwal, H. Venkannagari, L. Lehtiӧ, Structural basis of selective inhibition
of human tankyrases, J. Med. Chem. 55 (2012) 1360e1367.
[35] H. Gunaydin, Y. Gu, X. Huang, Novel binding mode of a potent and selective
tankyrase inhibitor, PLOS One 7 (2012) e33740.
A membrane-associated b-catenin/Oct4 complex correlates with ground-state
pluripotency in mouse embryonic stem cells, Development 140 (2013)
1171e1183.
[52] A. Ferrer-Vaquer, A. Piliszek, G. Tian, R.J. Aho, D. Dufort, A.-K. Hadjantonakis,
[36] H. Bregman, N. Chakka, A. Guzman-Perez, A.;H. Gunaydin, Y. Gu, X. Huang,
V. Berry, J. Liu, Y. Teffera, L. Huang, B. Egge, E.L. Mullady, S. Schneider,
P.S. Andrews, A. Mishra, J. Newcomb, R. Serafino, C.A. Strathdee, S.M. Turci,
C. Wilson, E.F. DiMauro, Discovery of novel, induced-pocket binding oxazo-
lidinones as potent, selective, and orally bioavailable tankyrase inhibitors,
J. Med. Chem. 56 (2013) 4320e4342.
A sensitive and bright single-cell resolution live imaging reporter of Wnt/b-
catenin signaling in the mouse, BMC Dev. Biol. 10 (2010) 121e129.
[53] J. Wray, T. Kalkan, S. Gomez-Lopez, D. Eckardt, A. Cook, R. Kemler, A. Smith,
Inhibition of glycogen synthase kinase-3 alleviates Tcf3 repression of the
pluripotency network and increases embryonic stem cell resistance to dif-
ferentiation, Nat. Cell Biol. 13 (2011) 838e845.
€
[37] T. Karlberg, N. Markova, I. Johansson, M. Hammarstrom, P. Schütz, J. Weigelt,
[54] N. Lyashenko, M. Winter, D. Migliorini, R. Biechele, R.T. Moon, C. Hartmann,
H. Schüler, Structural basis for the interaction between tankyrase-2 and a
potent Wnt-signaling inhibitor, J. Med. Chem. 53 (2010) 5352e5355.
Differential requirement for the dual functions of b-catenin in embryonic
stem cell self-renewal and germ layer formation, Nat. Cell Biol. 13 (2011)
753e761.
[38] T. Haikarainen, H. Venkannagari, M. Narwal, E. Obaji, H.-W. Lee, Y. Nkizinkiko,
€
L. Lehtio, Structural basis and selectivity of tankyrase inhibition by a Wnt
[55] W. Kabsch, Automatic processing of rotation diffraction data from crystals of
initially unknown symmetry and cell constants, J. Appl. Crystallogr. 26 (1993)
795e800.
[56] G.N. Murshudov, A.A. Vagin, E.J. Dodson, Refinement of macromolecular
structures by the maximum-likelihood method, Acta Crystallogr. D. Biol.
Crystallogr. D 53 (1997) 240e255.
signaling inhibitor WIKI4, PLoS One 8 (2013) e65404.
€
[39] L. Lehtio, R. Collins, S. van den Berg, A. Johansson, L.-G. Dahlgren,
€
M. Hammarstrom, T. Helleday, L. Holmberg-Schiavone, T. Karlberg, J. Weigelt,
Zinc binding catalytic domain of human tankyrase 1, J. Mol. Biol. 379 (2008)
136e145.
[40] A. Nathubhai, P.J. Wood, M.D. Lloyd, A.S. Thompson, M.D. Threadgill, Design
and discovery of 2-arylquinazolin-4-ones as potent and selective inhibitors of
tankyrases, ACS Med. Chem. Lett. 4 (2013) 1173e1177.
[41] Z. Hua, H. Bregman, J.L. Buchanan, N. Chakka, A. Guzman-Perez, H. Gunaydin,
X. Huang, Y. Gu, V. Berry, J. Liu, Y. Teffera, L. Huang, B. Egge, R. Emkey,
E.L. Mullady, S. Schneider, P.S. Andrews, L. Acquaviva, J. Dovey, A. Mishra,
J. Newcomb, D. Saffran, R. Serafino, C.A. Strathdee, S.M. Turci, M. Stanton,
C. Wilson, E.F. DiMauro, Development of novel dual binders as potent, se-
lective, and orally bioavailable tankyrase inhibitors, J. Med. Chem. 56 (2013)
10003e10015.
[57] P. Emsley, K. Cowtan, Coot: model-building tools for molecular graphics, Acta
Crystallogr, D. Biol. Crystallogr. D 60 (2004) 2126e2132.
[58] D. Zhao, Q. Shen, Y.-R. Zhou, J.-X. Li, KOtBu-mediated stereoselective addition
of quinazolines to alkynes under mild conditions, Org. Biomol. Chem. 11
(2013) 5908e5912.
[59] A. Couture, H. Cornet, P. Grandclaudon, An expeditious synthesis of 2-aryl-
and 2-alkylquinazolin-4(3H)-ones, Synthesis (1991) 1009e1010.
[60] R. Mekala, R. Akula, R.R. Kamaraju, C.K. Bannoth, S. Regati, J. Sarva, An efficient
synthesis of 2-substituted quinazolin-4(3H)-ones catalyzed by iron(III) chlo-
ride, Synlett (2014) 821e826.
[42] M.D. Shultz, A.K. Cheung, C.A. Kirby, B. Firestone, J. Fan, C.H.-T. Chen, Z. Chen,
D.N. Chin, L. DiPietro, A. Fazal, Y. Feng, P.D. Fortin, T. Gould, B. Lagu, H. Lei,
F. Lenoir, D. Majumdar, E. Ochala, M.G. Palermo, L. Pham, M. Pu, T. Smith,
[61] P. Mani Chandrika, T. Yakaiah, A. Raghu Ram Rao, B. Narsaiah, N. Chakra
Reddy, V. Sridhar, J. Venkateshwara Rao, Synthesis of novel 4,6-disubstituted
quinazoline derivatives, their anti-inflammatory and anti-cancer activity
(cytotoxic) against U937 leukemia cell lines, Eur. J. Med. Chem. 43 (2008)
846e852.
[62] H. Zhao, H. Hua, R. Qiao, Copper-catalyzed direct amination of ortho-func-
tionalized haloarenes with sodium azide as the amino source, J. Org. Chem. 75
(2010) 3311e3316.
ꢀ
T. Stams, R.C. Tomlinson, B.B. Toure, M. Visser, R.M. Wang, N.J. Waters,
W. Shao, Identification of NVP-TNKS656: the use of structure-efficiency re-
lationships to generate a highly potent, selective, and orally active tankyrase
inhibitor, J. Med. Chem. 56 (2013) 6495e6511.
ꢁ
ꢁ
[43] M. Tisler, B. Stanovnik, Z. Zrimsek, C. Stropnik, C. Heterocycles, CCI. A novel
synthesis of heterocyclic thioformylamines, Synthesis (1981) 299e300.