Inhibitor development of MTH1 via high-throughput screening with fragment based library and MTH1 substrate binding cavity

Bioorganic Chemistry 110 (2021) 104813

Contents lists available at ScienceDirect

Bioorganic Chemistry

journal homepage: www.elsevier.com/locate/bioorg

Inhibitor development of MTH1 via high-throughput screening with

fragment based library and MTH1 substrate binding cavity

Cheng Peng^a, Yu-Hsuan Li^b, Chao-Wu Yu^c, Ze-Hua Cheng^c, Jia-Rong Liu^c, Jui-Ling Hsu^d,

,

e

,c

Ling-Wei Hsin^c, Chen-Tsung Huang^e, Hsueh-Fen Juan^b^,f, Ji-Wang Chern^b

,

,b,f,*

Yi-Sheng Cheng^a

^aInstitute of Plant Biology, National Taiwan University, Taipei 10617, Taiwan

^bDepartment of Life Science, National Taiwan University, Taipei 10617, Taiwan

^cSchool of Pharmacy, National Taiwan University, Taipei 10050, Taiwan

^dDepartment of Pharmacy, New Taipei Municipal TuCheng Hospital, Chang Gung Memorial Hospital, New Taipei City 236017, Taiwan

^eGraduate Institute of Biomedical Electronics and Bioinformatics, National Taiwan University, Taipei 10617, Taiwan

^fGenome and Systems Biology Degree Program, Academia Sinica and National Taiwan University, Taipei 10617, Taiwan

A R T I C L E I N F O

A B S T R A C T

Keywords:

MutT Homolog 1 (MTH1) has been proven to hydrolyze oxidized nucleotide triphosphates during DNA repair. It

can prevent the incorporation of wrong nucleotides during DNA replication and mitigate cell apoptosis. In a

cancer cell, abundant reactive oxygen species can lead to substantial DNA damage and DNA mutations by base-

pairing mismatch. MTH1 could eliminate oxidized dNTP and prevent cancer cells from entering cell death.

Therefore, inhibition of MTH1 activity is considered to be an anti-cancer therapeutic target. In this study, high-

throughput screening techniques were combined with a fragment-based library containing 2,313 compounds,

which were used to screen for lead compounds with MTH1 inhibitor activity. Four compounds with MTH1 in-

hibitor ability were selected, and compound MI0639 was found to have the highest effective inhibition. To

discover the selectivity and speciﬁcity of this action, several derivatives based on the MTH1 and MI0639 complex

structure were synthesized. We compared 14 complex structures of MTH1 and the various compounds in com-

bination with enzymatic inhibition and thermodynamic analysis. Nanomolar-range IC₅₀inhibition abilities by

enzyme kinetics and K_dvalues by thermodynamic analysis were obtained for two compounds, named MI1020

and MI1024. Based on structural information and compound optimization, we aim to provide a strategy for the

development of MTH1 inhibitors with high selectivity and speciﬁcity.

Oxidative DNA damage

8-Oxo-dGTP

Ultra-high-throughput drug screening

Isothermal titration calorimetry

1. Introduction

[5,6]. When excess oxidized dNTPs form in cancer cells, high levels of

resultant DNA mispairing lead to programmed cell death.

In cancer cellular metabolism, a large amount of reactive oxygen

species (ROS) can accumulate in the process of abnormal proliferation

[1–3]. Excessive ROS causes the dNTP pool to form oxidized dNTPs such

as 8-oxo-dGTP and 2-OH-dATP. Oxygen free radicals oxidize the gua-

nine at the C-8 position and result in rotation of the chemical structure

from anti-form to syn-form [4]. Therefore, DNA mispairing, or so-called

DNA damage, occurs through changes from G–C pairing to A–T pairing

MutT homologue 1 (MTH1) is an emerging anti-cancer target that

has been shown to remove oxidized dNTPs [7–11]. The homologous

protein of MTH1 includes MTH2 and MTH3. MTH1 showed the highest

activity against 8-oxo-dGTP. MTH1 is able to make the hydrogen

bonding with 8-oxo-dGTP, but MTH2 binding pocket is poorly ﬁtting 8-

oxo-dGTP. MTH3 has far less activity than MTH1 against 8-oxo-dGTP.

Several studies using immunohistochemical staining have revealed

Abbreviations: MTH1, MutT Homolog 1; ROS, reactive oxygen species; uHTS, ultra high-throughput drug screening; IC₅₀, half maximal inhibitory concentration;

Ki, inhibitory constant; cLogP, partition coefﬁcient; ITC, isothermal titration calorimetry; K_d, binding constant; APBS, Adaptive Poisson–Boltzmann Solver software;

CASTp, Computed Atlas of Surface Topography of proteins software; LE, ligand efﬁciency; HA, heavy atom count; ΔG, Gibbs free energy of binding; ΔH, enthalpy

change; ΔS, entropy change; EEC, enthalpy–entropy compensation; B-factor, temperature factor.

* Corresponding author at: Department of Life Science, National Taiwan University, Taipei 10617, Taiwan.

E-mail address: chengys@ntu.edu.tw (Y.-S. Cheng).

https://doi.org/10.1016/j.bioorg.2021.104813

Received 5 January 2021; Received in revised form 25 February 2021; Accepted 6 March 2021

Available online 10 March 2021

Published by Elsevier Inc.

This is an open access article under the CC BY-NC-ND license

(http://creativecommons.org/licenses/by-nc-nd/4.0/).

C. Peng et al.

Bioorganic Chemistry 110 (2021) 104813

that MTH1 protein is expressed during growth of cancer cells and tissue

[12–15]. MTH1 is involved in colitis-associated carcinogenesis, esoph-

ageal squamous cell carcinoma, glioblastoma, and parotid adenoma

[15–18]. Rai et al. used shMTH1 xenograft to show that inhibition of

MTH1 can repress the growth of non-small-cell lung cancer in mice [19].

Overexpression of MTH1 can clean up oxidized dNTPs in oncogenic RAS

induction, and knockout of MTH1 causes accumulation of oxidized

dNTPs in cancer cells [20,21]. Furthermore, MTH1 plays an important

role in the RAS pathway by reducing internal oxidative pressure in

cancer cells [22,23]

we selected, the pocket changes are also not consistent and high ﬂexi-

bility. We think selectivity is an issue in the development of MTH1

inhibitors.

In this study, the ultra high-throughput drug screening (uHTS)

technique was combined with a fragment-based library containing

2,313 compounds that was used to screen for leading compounds with

inhibitory ability against MTH1. One of four compounds with the

highest inhibition ability, named MI0639 (IC₅₀of 1.5 μM), was selected

as the hit during compound optimization. Fourteen compounds based on

MI0639 were synthesized, and their inhibition abilities (IC₅₀) by

biochemical assay were tested and binding constants (K_d) were deter-

mined by thermodynamic analysis. The crystal structures of fourteen

compounds in complex with MTH1 were resolved to reveal the modes of

binding of the compounds. By combining structural information for

MTH1 inhibitors and their IC₅₀and K_dvalues, we provide a summary of

the selectivity of MTH1 toward inhibitors to provide information for

further inhibitor design and synthesis.

MTH1-deﬁcient mice have been generated, and the susceptibility of

mice to spontaneous tumorigenesis has been examined. The results show

no signiﬁcant difference in survival rate between MTH1+/+ and

MTH1ꢀ /ꢀ mice [24]. Knockdown of MTH1 expression by siRNA was

performed in another study using VH10 cell line, and the results also

showed no difference in cell survival rate [25]. These ﬁndings revealed

that MTH1 might not be essential in normal cells. However, MTH1 is

required in cancer cell survival based on a study of U2OS cells, in which

transfection with MTH1 siRNA resulted in decreased cell survival [26].

Based on the important role of MTH1 in cancer cells, MTH1 is a target for

the development of inhibitors for anti-cancer therapy [27]

2. Results and discussion

2.1. Chemistry

In 2011, the complex structure of MTH1 and 8-oxo-dGMP was re-

ported [28]. The complex structure showed MTH1 directly bound to 8-

oxo-dGMP in anti-conformation. Therefore, the oxidized form of dGTP,

8-oxo-dGTP, is a substrate of MTH1 and MTH1 can prevent 8-oxo-dGTP

binding to incorporate into DNA from mispairing during DNA replica-

tion. Two potent inhibitors, TH588 and TH287, as well as their complex

structures with MTH1, have been reported and analyzed [25]. TH588

has been shown to repress malignant melanomas in mice as well SW480

colorectal, glioblastoma, and MCF7 breast tumor growth [25,29,30].

TH287 has been shown to be more effective in MTH1 inhibition and to

repress oral cancer [31]. According to a research report by Kawamura

et al. (2016), the inhibition ability of TH588 may involve interference

with microtubule formation [32] and blockage of tubulin polymeriza-

tion during mitosis to produce more ROS, which then repress cancer

proliferation [33]. TH588 has been found to be an effective inhibitor in

animal trials, and TH588 analogs have been used in clinical trials

[34,35]. Recently, several studies have demonstrated that the function

and activity of MTH1 may be inﬂuenced by the development of different

inhibitors. In 2014, an inhibitor, crizotinib, was screened from a kinase

inhibitor collection found to inhibit dual c-MET/ALK kinases [36].

However, its enantiomer, S-crizotinib, also showed inhibitory ability

against MTH1 and could reduce tumor volume in a carcinoma xenograft

model, as well as suppress osteosarcoma [36,37]. Recently, a Toll-like

receptor 7 (TLR7) agonist with antitumor ability was found to be able

to interact with MTH1. The study designed and synthesized several

compounds based on the TLR7 agonist protein structure, and it obtained

three kinds of synthesized compounds showing MTH1 binding ability

and ﬁve additional complex structures of MTH1 compounds [38]. In

2018, Fredirk et al. reported a new, powerful MTH1 inhibitor, BAY-707

[39]. In 2019, 3-isomangostin was reported in a study of compound

cocktails, and its mode of binding to MTH1 was conﬁrmed using X-ray

crystallography [40]. However, not all synthesized MTH1 inhibitors

have been effective in preventing cell growth and in inducing apoptosis

[38,39]. Inhibitors from previous studies are known to exist and have

low selectivity for MTH1 (Fig. S1).

To obtain the target compounds, a series of reactions were carried

out as is shown in Scheme 1. The precursor compound 2 was prepared

through cyclization of commercially available diethyl 2-ethylmalonate

(1) and guanidine hydrochloride in the presence of sodium methoxide.

The second step was to transform the hydroxide moiety in compound 2

into chloride. Compound 4 was obtained by chlorination of compound 2

with Vilsmeier-Haack-Arnold reagent and immediate deprotection of

(dimethyl-amino) methylene group, which was carried out following the

procedure in literature [41]

The synthetic process for MI1001, MI1003 and MI1006 was shown

in Scheme 2. To introduce a methylamino group at the 2-position of

compound 4, it was treated with methylamine hydrochloride in the

presence of base with isopropyl alcohol as solvent, heated in sealed tube

to afford compound 5. Then, compound 5 was coupled with appropriate

amines in diphenyl ether, affording the target compounds MI1001,

MI1003 and MI1006. Due to the steric hindrance of ethyl group,

nucleophilic attack was not easy to occur at the 4-position of compound

5. Therefore, the last step was carried out under extremely high tem-

perature in sealed tube. And diphenyl ether was chosen as the solvent for

this reaction on account of its extra high boiling point.

In an attempt to synthesize 6-chloro-5-ethyl-N⁴-methylpyrimidine-

2,4-diamine (2), another approach using DMF as solvent was investi-

gated, affording compound 5 as the major compound in 36% yield.

Through this reaction, a minor product has been isolated and fully

characterized as 5-ethyl-N⁴,N⁶-dimethylpyrimidine-2,4,6-triamine (6).

Since the former approach mentioned in Scheme 2 has higher yield and

more convenient puriﬁcation process, this method was abandoned.

However, the minor compound (6) accidentally obtained from this re-

action was also considered as a potential target compound, and saved for

further evaluation.

To investigate the impact of cyclopropyl group on potency and

metabolic stability of drug, we synthesized a series compounds with

cyclopropyl group in replacement of methyl group. A cyclopropyl group

was introduced at the 4-position of compound 4 through nuchleophilic

substation of compound 4 and cyclopropyl amine. The reaction took

place in n-butanol with DIPEA as base under traditional heating condi-

tion. After that, compounds MI1013, MI1014 and MI1016 were gained

through the coupling of intermediate compound 7 with appropriate

amine. Compound 4 was treated with morpholine and DIPEA using

diphenyl ether as solvent and heated in sealed tube affording MI1013,

MI1014 and MI1016. Compounds MI1014 and MI1016 were prepared

through the same approach as for compound MI1013 with adjustments

in temperature, reacting time and puriﬁcation method. The synthetic

route of compounds MI1013, MI1014 and MI1016 was depicted in

The homologous protein of MTH1 includes MTH2 and MTH3. MTH1

showed the highest activity against 8-oxo-dGTP than MTH2 and MTH3.

For comparison structure with MTH1 and MTH2, MTH1 is able to make

the hydrogen bonding with 8-oxo-dGTP by helix α2 residues Asp119,

Asp120 and sheet β4 Asn33, but relative to MTH2 binding pocket is

Gly137, Leu138 and Gln44. This variance causes a distinct selectivity

that MTH2 binding pocket is poorly ﬁtting 8-oxo-dGTP [26]. This ex-

plains binding pocket is crucial. And due to development of inhibitors

for MTH1 in recent years, the inhibitors diversity implies that the pocket

of MTH1 can accommodate inhibitors by different shapes, as to the hits

2

C. Peng et al.

Bioorganic Chemistry 110 (2021) 104813

Scheme 1. Synthesis of compound 4^a. ^aReagents and conditions: (a) Guanidine hydrochloride, NaOMe, MeOH, 85 ^◦C, 2 h; (b) (Chloromethylene)dimethyliminium

chloride, CHCl₃, reﬂux, 4 h; (c) HCl, EtOH/H₂O, 50 ^◦C, 2 h.

Scheme 2. Synthesis of compounds MI1001, MI1003 and MI1006^a. ^aReagents and conditions: (a) Methylamine hydrochloride, DIPEA, IPA, 100 ^◦C, sealed tube, 17 h;

(b) Morpholine, DIPEA, diphenyl ether, 210 ^◦C, sealed tube, 17 h; (c) 1-Methylpiperazine, DIPEA, diphenyl ether, 240 ^◦C, sealed tube, 17 h; (d) Piperidine, DIPEA,

diphenyl ether, 240 ^◦C, sealed tube, 17 h.

Scheme 3.

In order to explore the effect of ethyl group in the structure on ac-

9 was coupled with appropriate amine through the same method as the

synthetic process of compound MI020 with modiﬁcation in tempera-

ture, reacting time and the way of work-up [42].

tivity, we also synthesized a series of compounds without the ethyl

group for comparison. The synthesis of N⁴-methyl-6-morpholinopyr-

imidine-2,4-diamine (MI1020) was carried out beginning with

commercially available 4,6-dichloropyrimidin-2-amine (8) as shown in

Scheme 4. Compound 8 was treated with methylamine hydrochloride in

presence of base, and heated in isopropyl alcohol to give 6-chloro-N⁴-

methylpyrimidine-2,4-diamine (9). A treatment of compound 9 with

morpholine and DIPEA using diphenyl ether as solvent in heating con-

dition afforded MI1020. To proceed structure activity relationship

(SAR) studies on MI1020, compounds MI1020, MI1012, MI1026 and

MI1030 were synthesized with different amino substitution on 6-posi-

tion of MI1020. To obtain MI1012, MI1026 and MI1030, compound

The synthetic process of compounds MI1025, MI1022, MI1024 and

MI1029 is shown as Scheme 5. To convert the chloro group at 4-position

of compound 8 into aminocyclopropyl group, compound 8 was treated

with cyclopropyl amine and DIPEA in n-butanol affording compound 10.

Compound 10 was then coupled with appropriate amines using the same

method as the synthetic process of compounds MI1020, MI1012,

MI1026 and MI1030 mentioned before, affording the target compounds

MI1025, MI1022, MI1024 and MI1029.

Scheme 3. Synthesis of compounds MI1013, MI1014 and MI1016^a. ^aReagents and conditions: (a) Cyclopropylamine, DIPEA, n-butanol, 95 ^◦C, 18 h; (b) Morpholine,

DIPEA, diphenyl ether, 210 ^◦C, sealed tube, 17 h; (c) 1-Methylpiperazine, DIPEA, diphenyl ether, 240 ^◦C, sealed tube, 17 h; (d) Piperidine, DIPEA, diphenyl ether,

240 ^◦C, sealed tube, 17 h.

3

C. Peng et al.

Bioorganic Chemistry 110 (2021) 104813

Scheme 4. Synthesis of compounds MI1020, MI1012, MI1026 and MI1030^a. ^aReagents and conditions: (a) Methylamine hydrochloride, DIPEA, IPA, 85 ^◦C, sealed

tube, 17 h; (b) Morpholine, DIPEA, diphenyl ether, 150 ^◦C, sealed tube, 17 h; (c) 1-Methylpiperazine, DIPEA, diphenyl ether, 150 ^◦C, sealed tube, 17 h; (d) Piperidine,

DIPEA, diphenyl ether, 150 ^◦C, sealed tube, 17 h; (e) 1-(oxetan-3-yl)piperazine, DIPEA, diphenyl ether, 150 ^◦C, sealed tube, 7 h.

Scheme 5. Synthesis of compounds MI1025, MI1022, MI1024 and MI1029^a. ^aReagents and conditions: (a) Cyclopropylamine, DIPEA, n-butanol, 95 ^◦C, 17 h; (b)

Morpholine, DIPEA, diphenyl ether, 150 ^◦C, sealed tube, 17 h; (c)1-Methylpiperazine, DIPEA, diphenyl ether, 150 ^◦C, sealed tube, 17 h; (d) Piperidine, DIPEA,

diphenyl ether, 150 ^◦C, sealed tube, 17 h; (e) 1-(oxetan-3-yl)piperazine, DIPEA, diphenyl ether, 150 ^◦C, sealed tube, 7 h.

2.2. A compound library via uHTS

2.3. Overview of identifying potential inhibitors

In order to obtain possible compounds with inhibitory ability against

MTH1, a library with 2,313 compounds from ChemBridge was screened

using a screening uHTS system. Recombinant MTH1 protein was ﬁrst

expressed in an Escherichia coli expression system and puriﬁed by fast

protein liquid chromatography (Supporting Information Fig. S2a and

S2b). The detection method was based on PPiLight™ assay and opti-

mized for uHTS. After screening, four out of 12 compounds, MI0639,

MI0320, MI0652, and MI0861, which had the highest inhibition ability

To reveal the modes of binding of the selected four compounds in the

MTH1 protein structure, we resolved these complex structures by X-ray

crystallography. After crystallization, the orthorhombic crystals were

used in the X-ray diffraction analysis (Supporting Information Fig. S2c).

MTH1 structure shows a typical Nudix motif with an α-helix on each side

and an intermediate series of several β-sheets connected by loops

(Fig. 1a). We obtained the crystals of MTH1 in complex with compounds

MI0639, MI0320, and MI0861 using the co-crystallization and soaking

method, and the resolutions after data collection ranged from 1.7 to 2.2

Å (Supporting Information Table S1). After model rebuilding, the com-

pounds were located at the binding pocket of MTH1 and mainly formed

hydrogen bonds between residues Asp119 and Asp120 in the second

(1–15

μM), were selected for description of structural differences

(Table 1). MI0639, which had the highest IC₅₀, contains a 2-aminopyr-

imidine substituent similar to those of TH588 and TH287 by coinci-

dence [25]. However, MI0639 has an additional ethyl substituent of 2-

aminopyrimidine and has a morpholinyl substituent in place of the

phenol ring of TH588. The morpholinyl substituent of MI0639 is less

ﬂattened than the phenol ring. The linking of 2-aminopyrimidine with a

ring substitution, like in MI0639 and MI0541, may give inhibitory

ability better than those of MI0231 and MI0609. This indicates that ring

labeling of 2-aminopyrimidine is a better direction for optimization. The

second compound, MI0320 with quinolone, is similar to AZ19 [38]; the

difference is that MI0320 lacks a phenol substituent and contains car-

bohydrazide. The IC₅₀of MI0232 is higher than that of MI0320, and the

three heterocyclic connections of MI0232 provide unfavorable inhibi-

tion. The third one, MI0652 with triazolo-pyrimidine, is similar to ess-

ramycin [43] and the IC₅₀of MI0652 is lower than those of MI0026 and

MI0114. A phenomenon that could explain this is that the 2-ethyl sub-

stituent of MI0026 and the location of the methylphenol substituent of

MI0114 reduce their binding ability. The fourth one, MI0861 with

quinolinedione, is a novel structure among the MTH1 inhibitors. In the

quinolinedione structure, modiﬁcation of the 4-phenol ring affects the

inhibition ability. The IC₅₀of MI0861 is better inhibition than that of

MI0700 and MI0195. This means that the methoxyphenol ring on qui-

nolinedione is better than the methyl or ﬂuorophenol rings.

α

-helix as well as between residues Asn33 and Gly34 in the third β-sheet.

For the compound MI0652, crystallization of MI0652 with MTH1 could

not be obtained because of incidental precipitation in the drops.

Here, we showed the binding modes of the compounds MI0639,

MI0320, and MI0861 in complex with MTH1. For MI0639, the residues

Asp119, Asp120, and Asn33 of MTH1 formed hydrogen bonds with the

N1, N2, and N3 position of the 2-aminopyrimidine moiety of MI0639

(Fig. 1b). Two hydrophobic interactions were formed by the residues

Phe72, Trp117, and Phe139 surrounding the methyl substituent at the C-

4 position of 2-aminopyrimidine, as well as the residues Phe27, Phe72,

Phe74, and Trp117 surrounding the ethyl substituent at the C-5 position

–

of the 2-aminopyrimidine of MI0639. A weak C H⋯O hydrogen bond

could be observed between the C-2 position of morpholine and the Thr8

backbone of the carboxyl substituent. Furthermore, the 2-aminopyrimi-

dine ring interacts with the sidechain of Trp117 by π–π stacking. For

MI0320, hydrogen bonds were formed between residue Asp120 and the

nitrogen of the carbohydrazide substituent, as well as between residue

Asp119 and the N4 position of quinoline. An additional hydrogen bond

was formed by the residue Gly34 backbone nitrogen and the oxygen of

the amino substituent (Fig. 1c). A hydrophobic interaction was formed

by the residues Phe27, Phe72, Phe74, and Trp117 surrounding the

heterocyclic ring of MI0320. For MI0861, the MTH1 inhibition ability

4

C. Peng et al.

Bioorganic Chemistry 110 (2021) 104813

Table 1

Chemical structures as well as biochemical and physicochemical data of selected MTH1 inhibitory compounds from uHTS.

Classiﬁcation

Group

Compound

ID

Potency

Physicochemical Properties

Structure

Biochemical IC₅₀

1.5

(μ

M)

Ki (

μ

M)

cLogP^a

MW (Da)

222

Solubility^b(mM)

7.65

2-Aminopyrimidine

MI0639

0.97

1.47

MI0541

7

4.4

2.28

216

6.79

MI0231

MI0609

11

16

7

3.13

1.71

294

167

0.29

11.2

10.1

Quinoline

MI0320

1.2

0.77

0.65

220

5.4

MI0232

MI0652

15

9.5

4.6

2.64

2.46

205

240

9.07

0.24

Triazolopyrimidine

7.2

MI0026

MI0114

MI0861

17

14

10.8

8.9

3.11

3.47

1.29

254

240

271

0.15

0.28

5.01

Quinolinedione

MI0700

MI0195

33

34

20.9

21.6

1.96

1.59

255

259

2.51

3.09

a

cLogP was calculated using ChemAxon Marvinsketch, version 19.1.

b

MW and solubility were estimated using Chemdraw Professional, version 16.0.

was weaker than that of the other selected compounds. After examina-

tion of the binding mode, MI0861 showed only one hydrogen bond

between Asp120 and the O1 position of quinolinedione (Fig. 1d). The

Computed Atlas of Surface Topography of proteins software (CASTp)

(Fig. 2) [46]. The shape of MTH1 resembles a glove, and the compounds

bind to the center site between the “thumb” and “index ﬁnger”. The

electrostatic surface in the binding pocket is contoured with a positive

charge from the residues Lys23, Arg31, and Lys38, although there are

two negatively charged residues, Asp119 and Asp120, at the bottom of

the pocket. After examination of the residues in the binding pocket,

more hydrogen bonds and hydrophobic interactions could be considered

for binding speciﬁcity. Furthermore, the calculated volumes of the

binding pocket of MTH1 are 640.3 Å³for apoMTH1, 714.7 Å³for MTH1

ꢀ MI0639, 673.8 Å³for MTH1 ꢀ MI0320, and 762.1 Å³for MTH1 ꢀ

MI0861 (Supporting Information Table S3). Interestingly, the volume

and IC₅₀values showed a correlation as MI0320 had the smallest volume

–

methoxy substituent also formed a weak C H⋯O hydrogen bond with

the oxygen atom of the Gly34 backbone carbonyl substituent. Hydro-

phobic interactions were formed by residues Leu9 and Val83 and the

methoxyphenyl substituent, as well as by residues Phe72 and Trp117

and the quinolinedione of MI0861. All hydrogen bonds are summarized

and listed in Table S2 in the Supporting Information.

To compare the differences in electrostatic charge and volumes of the

binding pocket of MTH1 before and after binding to compounds, we

calculated the molecular surface using Adaptive Poisson–Boltzmann

Solver software (APBS) [44,45] and calculated the volume using

5

C. Peng et al.

Bioorganic Chemistry 110 (2021) 104813

Fig. 1. Overall structure of MTH1 and MTH1 in

complex with selected inhibitors MI0639, MI0320,

and MI0861. (a) Ribbon diagram of MTH1 is color-

coded with α-helices in red, β-sheets in yellow, and

loops in teal. (b)–(d) Complex structures of MI0639,

MI0320, and MI0861 in MTH1 are shown with

electron density for 2Fo-Fc maps contoured at 1.0 σ.

The hydrogen bonds are shown as yellow dashed

lines between MTH1 and the inhibitors. (For inter-

pretation of the references to color in this ﬁgure

legend, the reader is referred to the web version of

this article.)

Fig. 2. Electrostatic surface of the MTH1–inhibitor complexes. The binding pocket of MTH1 in complex with MI0639 (a), MI0320 (b), and MI0861 (c). The hydrogen

bonds are shown as yellow dashed lines between MTH1 and the inhibitors, and the electrostatic surface is contoured from ꢀ 10 kT (red) to 10 kT (blue). (For

interpretation of the references to color in this ﬁgure legend, the reader is referred to the web version of this article.)

(673.8 Å³) and the highest IC₅₀(1.2

μ

M), while MI0861 had the largest

hydrogen bonds to MTH1 would be expected to have good binding

speciﬁcity and MTH1 selectivity.

volume (762.1 Å³) and the lowest IC₅₀(14

μM) (Table 1). According to

the analysis of volume and binding to the residues of the binding pocket,

modiﬁed compounds with better ﬁt into the binding pocket and more

To further examine the structural differences in the binding of in-

hibitors and substrate to MTH1, we superimposed the structures of

6

C. Peng et al.

Bioorganic Chemistry 110 (2021) 104813

MI0639, MI0320, and MI0861, the substrate 8-oxo-dGTP, and the pre-

viously reported inhibitor TH588 and AZ19 to reveal their binding

preferences. The overall structure of MI0639 is superimposed to that of

TH588 (Fig. 3a); however, TH588 contains an additional cyclopropyl

substituent as compared with MI0639. The orientation of the morpho-

linyl substituent of MI0639 is closed to that of the dichlorophenyl of

TH588 (Fig. 3a) and the 3-hydroxyl substituent on the deoxyribose of 8-

oxo-dGMP (Fig. 3b). Previous reports have demonstrated that colorectal

and breast tumors in mice are reduced after treatment with TH588

[25,29]. The results also demonstrated that TH588 is more slowly

metabolized in liver microsomes than TH287, maintaining high potency

in xenograft tumors in mice, and revealed that the cyclopropyl substit-

uent could maintain TH588 stability in vivo [25]. Therefore, the 4-

methyl substituent of the 2-aminopyrimidine of MI0639 could be

modiﬁed into an aminocyclopropyl substituent, or alternatively an N-

alkylation substituent, to increase the feasibility of animal trials. In

Fig. 3b, the Thr8 of MTH1 is shown forming hydrogen bonds with the 3-

hydroxyl substituent of 2-deoxyribose of 8-oxo-dGMP. By replacement

of the morpholinyl substituent by piperazine and replacement of the N

atom of piperazine by the hydroxyethyl substituent, hydrogen bonds

with oxygen of the carbonyl of Thr8 may be formed to enhance MTH1

binding ability. Previous reports have indicated that Met81 of MTH1

plays a role in binding selectivity for 8-oxo-dGTP and dGTP [28]. Here,

the ethyl substituent of the 2-aminopyrimidine of MI0639 near Met81 of

MTH1 could be considered to either be reserved or eliminated.

oxo-dGMP shows similar spatial arrangement (Fig. 3d). The quinoline-

dione ring of MI0861 and the guanine of 8-oxo-dGMP occupy the same

place, and the methoxy phenyl substituent of MI0861 coincides with the

location of the ribose of 8-oxo-dGMP. However, in the case of MI0861,

only a typical hydrogen bond is formed with the residue Asp120 of

MTH1. The fewer hydrogen bonds between MI0861 and MTH1 may lead

to weak MTH1 inhibitor activity. For further compound optimization

based on MI0861, a heterocyclic ring structure could be introduced to

increase the number of polar atoms available to form additional

hydrogen bonds with MTH1 residues, such as Thr8, Lys23, Asn33, and

Asp119 in MTH1.

2.4. MTH1 gene expression study

To generate a list of cancer types in which carcinogenesis is corre-

lated with MTH1 gene expression, we performed data analysis using the

Oncomine database of cancer microarray data [47]. The result shows

that MTH1 expression is signiﬁcantly higher in liver cancer and lung

adenocarcinoma than that of adjacent normal cells from cancer patients

(Fig. 4a). In addition, Kaplan–Meier analysis revealed that cancer pa-

tients with higher MTH1 expression had signiﬁcantly poorer survival

rates than those with lower MTH1 expression in both hepatocellular

carcinoma (Fig. 4b LIHC, p = 0.0043) and lung adenocarcinoma (Fig. 4b

LUAD, p = 0.0018). These results reveal that MTH1 expression is

correlated with the development of hepatocellular carcinoma and lung

adenocarcinoma.

In comparison with MI0320, AZ19 has an addition phenol ring

(Fig. 3c). The IC₅₀and K_dof AZ19 are about 10 nM [38], and its potency

is much higher than that of MI0320. This difference is not only due to the

ability of residue Asn33 to form a hydrogen bond with the aniline ni-

trogen of AZ19, but it also increases the hydrophobic interaction of the

residues Leu9 and Val83 with the phenol ring, which does not exist in

MI0320. In Fig. 2b, there is more space on the right side of MI0320, and

the phenol ring of AZ19 could occupy in the space of the binding cavity.

The result indicates that the better ﬁt into the binding pocket may

improve the binding ability. Structural comparison of MI0861 and 8-

2.5. Cell survival assay and ﬂow cytometric assay

To check the efﬁcacy of selected compounds in suppressing cancer

cell growth, two prostate cancer cell lines, PC-3 and DU145, were

treated with 20 μM compound for 48 hr and detected using sulforhod-

amine B staining. The results of this analysis showed that the compound

MI0639 reduced the proliferation rates of PC-3 cells by 26.7% and of

DU145 cells by 93.3%, but the other three compounds, MI0320,

Fig. 3. Structural comparison between selected in-

hibitors and inhibitors from the literature on MTH1.

The MTH1–compound complex structures are

superimposed over the inhibitors from the literature

and 8-oxo-dGMP, and represented as a cartoon

(MTH1) and sticks (inhibitors and 8-oxo-dGMP).

Close polar contacts are represented as dotted lines.

The dotted lines of the screening compounds

MI0639, MI0320, and MI0861 are yellow, and the

dotted lines of TH588 (PDB code:4N1U), AZ19 (PDB

code:5ANV), and 8-oxo-dGMP (PDB code:3ZR0) are

pink. (a) Superimposition of the MTH1-MI0639

(green) and MTH1-TH588 (cyan) complex struc-

tures, RMSD = 0.42 Å. (b) Superimposition of the

MTH1-MI0639 (green) and MTH1-8-oxo-dGMP

(purple) complex structures, RMSD = 0.26 Å. (c)

Superimposition of the MTH1-MI0320 (green) and

MTH1-AZ19 (yellow-orange) complex structures,

RMSD = 0.48 Å. (d) Superimposition of the MTH1-

MI0861 (green) and MTH1-8-oxo-dGMP (purple)

complex structures, RMSD = 0.34 Å. (For interpre-

tation of the references to color in this ﬁgure legend,

the reader is referred to the web version of this

article.)

7

C. Peng et al.

Bioorganic Chemistry 110 (2021) 104813

Fig. 4. High MTH1 gene expression is associated with poor survival rate in liver hepatocellular carcinoma and lung adenocarcinoma. (a) The Oncomine database

shows that the MTH1 (NUDT1) gene is upregulated in liver hepatocellular carcinoma (LIHC; n = 50) and lung adenocarcinoma (LUAD; n = 67) compared with

adjacent normal cells. The gene expression of MTH1 (NUDT1) was quantiﬁed and analyzed by paired Student t test: p < 0.01. (b) Kaplan–Meier survival analysis

between high MTH1 (NUDT1) gene expression and low expression of overall survival in patients with liver cancer (n = 149) and lung adenocarcinoma (n = 210): p

< 0.01.

MI0861, and MI0652, did not (Fig. 5). In addition, MI0639 showed an

IC₅₀of 6.5 ± 0.1 μM in repressing the proliferation of DU145 prostate

cancer cells (Fig. S3). The other three compounds, MI0320, MI0861, and

MI0652, showed lower repression rates, which might be due to poor

absorption by cancer cells. According to the cancer cell inhibition

experiment, MI0639 possessed efﬁcacy against MTH1 not only for in

vitro repression of enzyme activity but also for inhibition of proliferation

of prostate cancer cells.

We also investigated the effects of MI0639 on the proliferation, cell

cycle regulation with different concentrations 1, 3, 10 and 20 mM for 24

hr and 48 hr by ﬂow cytometry. The data demonstrated that the expo-

sure of DU145 cells to MI0639 for 24 hr induced a concentration-

dependent increase of cell population at G0/G1 phase of the cell cycle

(37.4%, 41.1%, 43.3%, 46.4% and 45.2% at 0, 1, 3, 10 and 20 mM

MI0639, respectively) and a profound increase at sub-G1 (5.3%, 5.3%,

5.4%, 14.3% and 27.6% at 0, 1, 3, 10 and 20 mM MI0639, respectively)

after a 48 hr treatment using ﬂow cytometric analysis of PI (Fig. S4). The

data substantiated that MI0639 induced an inhibitory effect of cell

proliferation and followed by an increase of cell death with DNA

Fig. 5. Inhibition of MTH1 reduces survival in cancer cells. The 20 μM MTH1

inhibitors were added to prostate cells PC-3 and DU145 for 48 hr. The cells

were ﬁxed and stained with sulforhodamine B, and the data were analyzed. The

IC₅₀for inhibition of cell proliferation based on MI0639 treatment of DU145

cells was 6.5 ± 0.1 μM.

8

C. Peng et al.

Bioorganic Chemistry 110 (2021) 104813

fragmentation (sub-G1 population). Therefore, the compound MI0639

would be the next priority for enhancement of binding speciﬁcity by

chemical optimization.

N-cyclopropyl substituent would be expected to enable additional

hydrogen bonds with Asp119. Finally, the oxygen atom of the mor-

pholine of MI0639 is close to Tyr7 and Thr8. To increase interaction

between Tyr7 and Thr8, we replaced the oxygen atom of morpholine

with carbon and nitrogen atoms to examine which atom is better for

optimization. In addition, the oxetanyl substituent was following to

connect the nitrogen of piperazine to form additional hydrogen bonds

with Thr8. Consequently, 14 compounds were synthesized, and their

biochemical and physiochemical properties were analyzed. They are

symbolized by an R1 substituent on the C-4 position and an R2 sub-

stituent on the C-6 position of 2-aminopyrimidine. All compounds are

listed in Table 2 for scaffold 1 with 5-ethyl substituent and Table 3 for

scaffold 2 without 5-ethyl substituent.

2.6. Chemical optimization and kinetic inhibition of derivatives

Since MI0639 showed effective inhibition ability for the cancer cell

line DU145, further chemical structure optimization was carried out.

Three parts of MI0639 were considered: (i) eliminating the 5-ethyl

substituent, (ii) replacing the 4-methyl substituent of 2-aminopyrimi-

dine with an N-methyl or N-cyclopropyl substituent, and (iii) replac-

ing the oxygen atom of morpholine with other atoms or an additional

alkyl substituent. In the structural comparison of MI0639, TH588, and 8-

oxo-dGMP, the 5-ethyl substituent of MI0639 is close to the Met81 and

Phe27 of MTH1. Therefore, the following compounds could be ﬁrst

divided into two templates, named scaffold 1 and scaffold 2, which

respectively contained and did not contain the 5-ethyl substituent.

Secondly, the N-cyclopropyl substituent of TH588 and the 6-hydroxy

substituent of 8-oxo-dGMP are oriented toward to the Asp119 residue.

Replacement of the 4-methyl substituent of MI0639 with an N-methyl or

The fourteen synthetic compounds could be further divided into

three groups after comparison with the IC₅₀of MI0639. The ﬁrst group,

with the highest IC₅₀inhibition ability of <41 nM, included four com-

pounds: MI1020, MI1024, MI1025, and MI1026. With an IC₅₀6.3 nM,

MI1024 showed the highest inhibition ability; its structure contained an

N-cyclopropyl substituent at R1 and a piperidine at R2. With a cLogP =

1.99, MI1024 was more lipophilic than the other three compounds in the

Table 2

Chemical structures, as well as biochemical and physicochemical data, of scaffold 1 compounds with 5-ethyl substituent based on the hit MI0639.

Compound

Potency

Physicochemical Properties

ID

Structure

R1

R2

Biochemical IC₅₀

6.3

(μM)

Ki (

μ

M)

cLogP^a

1.47

MW (Da)

222.29

Solubility^b(mM)

7.65

MI0639 (Hit)

4

MI1001

MI1003

57.3

26.1

36.2

165

1.41

1.47

237.31

250.35

3.81

3.09

MI1006

MI1013

71.1

45

2.48

1.87

235.34

263.35

0.47

1.14

263.9

167

MI1014

>1000

675

3

1.94

2.94

276.39

261.37

0.93

0.14

MI1016

4.8

a

b

cLogP was calculated using ChemAxon Marvinsketch, version 19.1.

MW and solubility were estimated using Chemdraw Professional, version 16.0.

9

C. Peng et al.

Bioorganic Chemistry 110 (2021) 104813

Table 3

Chemical structures as well as biochemical and physicochemical data of scaffold 2 compounds without 5-ethyl substituent based on the hit MI0639.

Compound

Potency

Physicochemical Properties

ID

Structure

R1

R2

Biochemical IC₅₀

6.3

(μM)

Ki (

4

μM)

LE

cLogP^a

1.47

MW (Da)

222.29

Solubility^b(mM)

7.65

MI0639 (Hit)

0.57

MI1012

MI1020

0.31

0.01

0.19

0.67

0.75

0.52

0.45

222.3

11.42

14.12

0.006

0.94

209.25

MI1022

MI1024

MI1025

1.5

0.54

0.68

0.65

0.98

1.99

0.92

248.33

233.32

235.29

3.43

0.52

4.2

0.006

0.041

0.004

0.025

MI1026

MI1029

0.022

2.0

0.014

1.28

0.77

0.46

1.52

0.82

207.28

290.37

1.76

3.19

MI1030

0.14

0.088

0.56

0.36

264.33

10.33

a

cLogP was calculated using ChemAxon Marvinsketch, version 19.1.

b

MW and solubility were estimated using Chemdraw Professional, version 16.0.

ﬁrst group. The compound MI1020, with an IC₅₀of 10 nM, contained an

N-methyl substituent at R1 and a morpholine at R2. MI1020 also showed

excellent inhibitory activity against MTH1, as well as high ligand efﬁ-

ciency, LE (LE = 0.75 kcal mol^{ꢀ 1}HA^{ꢀ 1}, calculated as LE = ΔG/HA) [48].

According to the solubility of the ﬁrst group, the N-methyl substituent at

R1 caused MI1020 to be more hydrophilic than MI1025, and MI1026

was more hydrophilic than MI1024 (Table 3). The second group, with

compounds: MI1012, MI1022, MI1029, and MI1030. The compound

MI1030 had the highest inhibitor activity in the second group with an

IC₅₀of 0.14 μM. The solubilities of MI1029 and MI1030, both carrying

oxetanyl piperazine at R2, were higher than those of MI1012 and

MI1022, which instead carried methyl piperazine. These results reveal

that inclusion of the oxetanyl substituent could increase water solubility

and molecular polarity. The third group, which was composed of weaker

inhibitors and compounds with no inhibitory ability, included six

moderate inhibitor ability (0.14–2

μM), was composed of four

10

C. Peng et al.

Bioorganic Chemistry 110 (2021) 104813

compounds: MI1001, MI1003, MI1006, MI1013, MI1014, and MI1016.

2.7. Thermodynamic analysis

The IC₅₀values of the third group ranged from 4.8 μM to 1 mM. Scaffold

1, in which the ethyl substituent belonged to the third group, showed the

worst inhibition of MTH1. Therefore, the ethyl substituent at C-5 posi-

tion is not considered to be conducive in molecular optimization. In

scaffold 2, the N-methyl substituent on R1 results in an IC₅₀value better

than that of the N-cyclopropyl substituent. Therefore, the cyclopropyl

substituent on R1 could slightly reduce inhibition on IC₅₀and might not

be useful in optimization. As shown in Table 3, the N-cyclopropyl sub-

stituent on R1 could increase lipophilicity and the oxygen-containing

substituent on R2 could increase solubility. In a comprehensive com-

parison of IC₅₀, a compound with the N-cyclopropyl substituent on R1

and the piperidine or piperazine on R2 would have better IC₅₀values

than a compound labeled with morpholine or piperazine linking the

oxetanyl ring on R2. Conversely, a compound with the N-methyl sub-

stituent on R1 and the morpholine or piperazine linking oxetanyl ring on

R2 would have better IC₅₀values than a compound labeled with

piperidine or piperazine on R2. The pairing of R1 and R2 with different

substituents revealed an interesting phenomenon, which is discussed in

the next section.

To reveal the binding modes between the derivatives and MTH1

protein, we performed thermodynamic analysis to determine the Gibbs

free energy of binding (ΔG), enthalpy change (ΔH), and entropy change

(ΔS) [49]. Since the enthalpy change (ΔH) could be determined from the

formation on the hydrogen bonds and the van der Waals force, and the

temperature-dependent entropy change (ꢀ TΔS) could be inﬂuenced by

hydrophobic interactions and the conformational change in space

[50,51], we would like to describe the thermodynamic change of MTH1

in binding to the derived compound. Negative ΔH and ꢀ TΔS values

mean that the molecular binding is stable and favorable. In the ﬁrst

group with the highest IC₅₀, the ΔG of MI1020, MI1024, MI1025, and

MI1026 are about ꢀ 11 kcal/mol (Fig. 6), and the K_dvalues are lower

than 10 nM. The results showed that the ﬁrst group has the best binding

constant against MTH1, and MI1024 still has the best ΔG of ꢀ 11.69 ±

0.07 kcal/mol and K_dof 2.9 ± 0.1 nM. We further compared the

chemical structures of MI1020 and MI1025 with MI0639. The results of

this comparison showed that the removal of the ethyl substituent from

–

MI0639 to replace CH₃on 2-aminopyrimidine of MI0639 by the N-

methyl (MI1020), as well as the N-cyclopropyl (MI1025) substituent,

could enhance ΔG and improve ΔH. This substituent replacement

Fig. 6. Thermodynamic proﬁles of the series of

optimized compounds binding to MTH1 based on ITC

experiments. Experimentally derived parameters

include the Gibbs free energy of binding (ΔG, blue

bar), enthalpy change (ΔH, green bar), and

temperature-dependent entropy change (ꢀ TΔS, red

bar). The presented data are mean values from at

least two titrations with calculated standard devia-

tion. (For interpretation of the references to color in

this ﬁgure legend, the reader is referred to the web

version of this article.)

11

C. Peng et al.

Bioorganic Chemistry 110 (2021) 104813

enabled the formation of a hydrogen bond between the Asp119 carboxyl

and N4 of the 2-aminopyrimidine motif of MI1020 and MI1025 but not

MI0639. The Asn33 sidechain carbonyl substituent also formed a

hydrogen bond with N2 of the 2-aminopyrimidine motif of MI1020 and

MI1025 (Fig. 7a) but not MI0639. Evidently, formation of these

hydrogen bonds leads to elevated ΔH. In comparison with the

isothermal titration calorimetry (ITC) results for MI1020 and MI1025,

we found that the additional cyclopropyl substituent of MI1025 effec-

tively improved ꢀ TΔS and slightly reduced ΔH. The Phe27 and Phe74

residues of MTH1 formed extra hydrophobic interactions with the

cyclopropyl substituent of MI1025 and could improve ꢀ TΔS (Fig. 7a).

This situation could be also found in the comparison of MI1024 with

MI1026. Moreover, ΔH could be maintained by the residue Tyr7 of

MTH1 to form additional hydrophobic interactions with the piperidine

2.8. Structure analysis

The removal of the 5-ethyl substituent of MI0639 and the replace-

ment of the 4-methyl substituent of the 2-aminopyrimidine with an N-

methyl or N-cyclopropyl substituent would cause the N4 atom of scaf-

fold 2 to form a stable hydrogen bond with the carboxyl substituent of

Asp119 (Fig. 7a) and increase the IC₅₀of inhibition ability against MTH1

to one order higher than that of MI0639. Previous studies have indicated

the advantage of cyclopropyl substituents in compounds that are not

easily metabolized by the liver [25,53,54]. The cyclopropyl substituent

on 2-aminopyrimidine in both MI1022 and MI1025 could induce more

ﬁtting by hydrophobic interactions with the periphery residues Phe72,

Phe74, Phe139, and Phe27 (Fig. 7a and c). Nevertheless, weaker IC₅₀

values and a reduced ΔH in the ITC experiment was apparent in the

cyclopropyl substituent of MI1022 and MI1025 in comparison with the

N-methyl substituent of MI1012 and MI1020 (Fig. 6). These results

indicated that the cyclopropyl substituent was not suitable in our com-

pound optimization.

–

of MI1024 and the sulfur atom of Met81 to form C H⋯O weak

hydrogen bond with the C-2 position of the piperidine of MI1024

(Fig. 7b). These results could compensate the ΔH reduction for the

cyclopropyl substituent effect of MI1024 and MI1025.

In the second group with the medium IC₅₀, the change in ΔH and

ꢀ TΔS also showed similar trends between MI1012 and MI1022, as well

as between MI1020 and MI1025 without and with the cyclopropyl

substituent on 2-aminopyrimidine motif. The cyclopropyl substituent in

both MI1022 and MI1025 could lead to additional hydrophobic inter-

action with Phe74, Trp117, and Phe139 of MTH1 (Fig. 7c and d),

thereby improving ꢀ TΔS and reducing ΔH. In comparison with MI1029

and MI1030, the cyclopropyl substituent of MI1029 could form addi-

tional hydrophobic interactions with Phe27, Phe74, and Trp117 of

MTH1 (Fig. 7d). Additionally, ꢀ TΔS of MI1029 (ꢀ 1.8 ± 1.7 kcal/mol)

and that of MI1030 (1.0 ± 0.2 kcal/mol) were favorably increased by the

oxetanyl piperazine. The correlation pattern of ꢀ TΔS and ΔH from

MI1012/MI1022 to MI1030/MI1029 corresponds to enthalpy–entropy

compensation (EEC) [52]. However, the oxetanyl piperazine of MI1030/

MI1029 oriented toward the entrance of the binding pocket showed

EEC. The reason for this is explained in the following section.

On the basis of the IC₅₀data, the compounds with 5-ethyl sub-

stituents on the 2-aminopyrimidine motif would reduce the inhibition

potency by over one order as compared with that of the compounds

without 5-ethyl substituents such MI1006/MI1026, MI1014/MI1022,

and MI1016/MI1024. The ethyl substituents in MI1006, MI1014, and

MI1016 would result in rotation of the R2 substituent (Fig. 8). The

rotational angle and temperature factor (B factor) are listed in Table 4.

The rotational angles of the R2 substituents are 60.2^◦for MI1006/

MI1026, 58.7^◦for MI1014/MI1022, and 59.7^◦for MI1016/MI1024,

which were affected by the ethyl substituent. Moreover, the B factors on

the R1 and R2 position of the compounds with ethyl substituents were

higher than those of the compounds without the ethyl substituent, such

as MI1006/MI1026, MI1014/MI1022, and MI1016/MI1024. The N-

cyclopropyl substituents of the compounds MI1006, MI1014, and

MI1016 were shown to have incomplete electron density maps as well as

higher temperature factors (Fig. 9). These results indicate that existence

of the ethyl, N-methyl, and N-cyclopropyl substituents simultaneously

Fig. 7. Binding of the scaffold 2 compounds to

MTH1. In scaffold 2 compounds, the R2 substituents

were lifted up by coexistence with the cyclopropyl

substituent (R1). (a) Close-up view of the super-

imposed compounds MI1020 (yellow orange) and

MI1025 (green), RMSD = 0.32 Å. (b) Close-up view

of the superimposed compounds MI1024 (green) and

MI1026 (light pink), RMSD = 0.28 Å. (c) Close-up

view of the superimposed compounds MI1012

(green) and MI1022 (cyan), RMSD = 0.18 Å. (d)

Close-up view of the superimposed compounds

MI1029 (green) and MI1030 (light orange), RMSD =

0.24 Å. The active-site residues and the compounds

are represented as sticks. The hydrogen bonds be-

tween the compounds MI1020 (a), MI1026 (b),

MI1012 (c), and MI1030 (d) with the N-methyl sub-

stituent and MTH1 are represented by the yellow

dotted lines. The hydrogen bonds between the com-

pounds MI1025 (a), MI1024 (b), MI1022 (c), and

MI1029 (d) with the cyclopropyl substituent and

MTH1 are represented by the pink dotted lines. The

carbon hydrogen bonds between MI1020 or MI1025

and Thr8 (a), MI1024 and Met81 (b), and MI1030

and Gly36 (d) are represented by cyan dotted lines,

and the hydrophobic interaction between MI1024

and Tyr7 (b) is shown by marine dotted lines. (For

interpretation of the references to color in this ﬁgure

legend, the reader is referred to the web version of

this article.)

12

C. Peng et al.

Bioorganic Chemistry 110 (2021) 104813

Fig. 8. Binding of the scaffold 1 compounds with ethyl substituent to MTH1. In the scaffold 1 compounds, the ethyl substituent was a factor of steric hindrance and

caused unfavored binding to MTH1. (a) Close-up view of the superimposed compounds MI1006 (green) and MI1026 (cyan), RMSD = 0.23 Å. (b) Close-up view of the

superimposed compounds MI1014 (green) and MI1022 (light pink), RMSD = 0.24 Å. (c) Close-up view of the superimposed compounds MI1016 (cyan) and MI1024

(yellow orange), RMSD = 0.27 Å. The active-site residues of MTH1 and the compounds are represented as sticks. The hydrogen bonds are represented as dotted lines.

The dotted lines between the compounds MI1006, MI1014, and MI1016 with the ethyl substituents and MTH1 are colored in yellow. The dotted lines between the

compounds MI1022, MI1024, and MI1026 without the ethyl substituent are colored in red. (For interpretation of the references to color in this ﬁgure legend, the

reader is referred to the web version of this article.)

resulted in steric hindrance and instability of the hydrogen bonds. In

Table 4

addition, the N-methyl and the N-cyclopropyl substituents showed

B-factor of inhibitors in the different crystal structures.

twisting in the binding pocket and indicated that the ethyl substituent

Inhibitor

Total (Å²)

R1 (Å²)

R2 (Å²)

Angle of R2 Rotation (Degree)

should not exist with the N-methyl and N-cyclopropyl substituents at the

same time. Consequently, the ethyl substituent was not considered

conducive for further optimization of compounds.

MI0639*

MI1020

MI1006*

MI1026

MI1014*

MI1022

MI1016*

MI1024

14.0

9.4

12.4

8.0

16.7

11.5

39.4

44.2

63.9

21.7

27.6

19.0

33.4^◦

0^◦

38.2

34.5

45.8

20.3

29.4

18.7

63.5

27.0

64.8

19.0

56.9

21.7

60.2^◦

0^◦

In scaffold 2 without the ethyl substituent, the compounds MI1022,

MI1024, MI1025, and MI1029 carrying N-cyclopropyl substituent at R1

showed a raised cyclic R2 substituent with N-methyl piperazine, piper-

idine, morpholine, and oxetanyl piperazine (Fig. 7). The K_d, IC₅₀, and

ΔG values of MI1022 and MI1029 were higher than those of MI1024 and

MI1025. The results reveal that MI1024 and MI1025 bind to MTH1 with

higher binding ability (Table 3 and Fig. 6). In contrast, the compounds

MI1012 (N-methyl piperazine), MI1020 (morpholinyl), MI1026 (piper-

idine), and MI1030 (oxetanyl piperazine) carrying the N-methyl

58.7^◦

0^◦

59.7^◦

0^◦

*

Indicates compound labeling with ethyl substituent on the 2-

aminopyrimidine.

Fig. 9. Electron density map and temperature factors

(B-factors) of the compounds MI1006, MI1014, and

MI1016 with the ethyl substituent. For the ethyl

substituent in the scaffold 1 compounds, the electron

density map and B-factors revealed that R1 and R2

were affected and caused incomplete mapping of R1,

as indicated by the arrowhead. (a) Electron density

map of MI1006 with the ethyl substituent super-

imposed with MI1026 in light sticks. (b) Electron

density map of MI1014 with the ethyl substituent

and MI1022 in light sticks. (c) Electron density map

of MI1016 with the ethyl substituent and MI1024 in

light sticks. The 2Fo-Fc electron density map of

compounds contoured at 1.0

σ level is colored in

gray. The bar of B factor ranges from low in blue to

high in red. Values of the B factor are listed in

Table 4. (For interpretation of the references to color

in this ﬁgure legend, the reader is referred to the web

version of this article.)

13

C. Peng et al.

Bioorganic Chemistry 110 (2021) 104813

substituent at R1 have similar ΔG values (ꢀ 10.8 to 11.69 kcal/mol)

(Fig. 6); the K_dand IC₅₀values of MI1012 and MI1030 are higher than

those of MI1020 and MI1026. The results also demonstrated that

MI1020 and MI1026 bind to MTH1 with higher binding ability. In

addition, there is a trend in the pairing of R1 and R2 in which the N-

methyl substituent is suitable for pairing with the polar substituent in

morpholine or oxetanyl piperazine, and the N-cyclopropyl substituent is

suitable for pairing with the nonpolar substituent in piperazine.

the N-methyl (R1) substituent of MI1020 (Fig. 7a). The result also pro-

vided an explanation as to why the IC₅₀of MI1020 is better than that of

MI1025 and MI1026. The IC₅₀value of MI1024 is also the lowest as

compared with MI1025 and MI1026 because the N-cyclopropyl (R1)

substituent could raise the piperidine of MI1024 in space and lead to the

–

formation of a weak C H⋯O hydrogen bond with Met81 at the C-2

position of piperidine. Bonding could not be observed in the piperidine

of MI1026 without the N-cyclopropyl substituent (Fig. 7b). Moreover,

the piperidine of MI1024 could form a hydrophobic interaction with the

Tyr7 side chain, which could not be observed with the morpholine of

MI1025. This might be due to the reduced hydrophobicity of the oxygen

atom of morpholine compared with that of piperidine. These results also

reveal why the IC₅₀of MI1024 is better than those of MI1025 and

MI1026. The compounds MI1020 and MI1024 had the best potency

among our optimization compounds.

Interestingly, the compounds MI1012 and MI1022 with the methyl

piperazine substituents at R2 were unable to form a stable hydrophobic

interaction with Tyr7, unlike the compounds MI1024 and MI1026 that

had the piperidine substituent. Since the N4 atom in piperazine

belonged to the sp³orbital with a lone pair of electrons, the hydro-

phobicity was insufﬁcient compared with that of piperidine. The extra

methyl substituent of piperazine reduced the polarity compared with

–

that of morpholine, which could form a C H⋯O weak hydrogen bond

with the Thr8 backbone. Therefore, the methyl piperazine of compounds

MI1012 and MI1022 reduced the inhibitory effect more greatly than did

MI1020, MI1024, MI1025, and MI1026 (Fig. 10a and b).

3. Conclusions

This study reported several compounds identiﬁed by uHTS, and it

further designed and optimized compounds according to the binding

pocket of MTH1 by combining enzyme kinetics, thermodynamics, and

protein structural information. The compound MI0639, which had the

greatest ability to bind to MTH1 has been identiﬁed. In further optimi-

zation, we found that the ethyl substituent and the N-methyl or N-

cyclopropyl substituent were not suitable for coexistence because the

ethyl substituent interferes with the hydrogen bonding between the

optimized compounds and Asp119 of MTH1. The unfavorable results

could be observed in the compounds of scaffold 1 (Table 1). However,

the ethyl substituent in MI0639 is favorable on the basis of thermody-

namic data and did not show steric hindrance such as interference with

the hydrogen bonding of Asp119. The complex structure of MI0639 and

MTH1 suggests that the ethyl substituent of MI0639 could also form

hydrophobic interactions with Phe27, Phe72, Phe74, and Trp117 of

MTH1.

As for the long oxetanyl piperazine substituent at R2 of the com-

–

pound MI1030, the piperazine formed a C H⋯O weak hydrogen bond

with the Gly36 backbone, and the IC₅₀of MI1030 was slightly better

than that of MI1012. The addition of the oxetanyl ring in MI1030 was

more ﬁtting to the binding pocket of MTH1, which prevented the

compounds MI1029 and MI1030 from easily oscillating in space. The

results correspond to why the ꢀ TΔS value of MI1029 and MI1030 is

lower and quite favorable in comparison to others. However, the

modiﬁcation of oxetanyl piperazine was close to the opening of the

pocket and might interact with the peripheral water molecules, resulting

in reduced IC₅₀and ΔH of MI1029 and MI1030 (Fig. 10c and d).

To compare the morpholine and piperidine substituents in the R2

substituent, the morpholine C-2 of MI1020 was capable of forming a

–

weak C H⋯O hydrogen bond with the carboxyl substituent of Thr8, but

the piperidine of MI1026 was not (Fig. 7a and b). The difference was just

the replacement of the C-4 of piperidine in MI1020 by the oxygen atom

of morpholine in MI1026. As for the morpholine substituent in MI1020

and MI1025, the hydrophobicity of the N-cyclopropyl (R1) substituent

of MI1025 was higher than that of MI1020 and was incompatible with

In scaffold 2, the methyl substituent (R1) of MI0639 was replaced by

the N-methyl or the N-cyclopropyl substituent, which enabled the for-

mation of an additional hydrogen bond between the optimized com-

pounds and MTH1. However, the cyclopropyl substituent (R1) in the

Fig. 10. Structural comparison of scaffold

2

compounds in R2 substituents. (a) Close-up view

of the superimposed compounds MI1012 (light or-

ange), MI1020 (yellow), and MI1026 (cyan). (b)

Close-up view of the superimposed compounds

MI1022 (light orange), MI1024 (cyan), and MI1025

(yellow). (c) Close-up view of the superimposed

compounds MI1012 (cyan) and MI1030 (yellow),

RMSD = 0.12 Å. (d) Close-up view of the super-

imposed compounds MI1022 (cyan) and MI1029

(yellow), RMSD = 0.18 Å. The active-site residues

and the compounds are shown as sticks, and the

binding pocket of MTH1, which resembles a glove, is

contoured in gray. (For interpretation of the refer-

ences to color in this ﬁgure legend, the reader is

referred to the web version of this article.)

14

C. Peng et al.

Bioorganic Chemistry 110 (2021) 104813

optimized compounds could slightly weaken the hydrogen bond

strength because of its hydrophobicity and then affect the R2 substitu-

ent, which prefers to be substituted by a six-membered ring. Therefore,

two optimized compounds MI1020 with morpholine and MI1024 with

piperidine have been validated to have the best IC₅₀and K_dvalues via

different binding modes. Compound MI1020 was shown to be a potent

inhibitor that competes with the substrate 8-oxo-dGTP. The morpholine

could form hydrogen bonds with Thr8 of MTH1. For compound MI1024,

the piperidine could interact with the Met81 sulfur atom. From our

analysis, the factors of improving selectivity and speciﬁcity include the

hydrogen bonding increasing, the regulation of R2 substituent polarity,

and additionally, preventing steric hindrance between the ethyl sub-

stituent and R1 substituent can enhance binding pocket ﬁtting. We

expect to design more speciﬁc and selective inhibitors against MTH1 to

avoid off-target effects.

UV detection at 254 nm, the ﬂow rate is 1.5 mL/min with using a solvent

gradient system, where mobile phase A = 0.1% Triﬂuoroacetic acid

(TFA) in water; B = Acetonitrile in 0.1% TFA solution. For MI1016,

MI1003 and MI1013, HPLC was obtained using Waters 717. The column

is XBridge BEH C18 Column (5 µm, 4.6 mm × 250 mm) with UV

detection at 220 nm, the ﬂow rate is 1.0 mL/min with using a solvent

gradient system, where mobile phase A = 100% Acetonitrile; B = 0.1%

TFA solution. For MI1030, MI1001, MI1012 and MI1006, UPLC was

obtained using Waters Acquity. The column is HSS T3 Column (1.8 µm,

2.1 × 50 mm) with UV detection at 220 nm, the ﬂow rate is 0.6 mL/min

with using a solvent gradient system, where mobile phase A = 0.1 N

NH₄HCO₃; B = Acetonitrile. All compounds purity was conﬁrmed to

have >95% purity. The data are supplied in the Supporting Information.

4.1.1. 2-Amino-5-ethylpyrimidine-4,6-diol (2)

Previous studies have shown that high expression of MTH1 can be

found in the prostate, liver, and lung cancer cell lines [55–58]. Ac-

cording to the Kaplan–Meier survival analysis, the overall ﬁve-year

survival rate is higher in liver and lung cancer patients with low

MTH1 expression than those with high MTH1 expression (Fig. 4).

Therefore, inhibitor development would be effective in such cancers.

Gad et al. reported that prostate cancer cells could be repressed by using

siRNA to reduce MTH1 expression [25]. In our study, prostate cancer

cell growth could also be repressed by treatment with MI0639. The in-

hibitors TH588 and S-crizotinib have been used to treat patients with

lung cancer [38,59]. As for liver cancer, MTH1-deﬁcient mice showed

spontaneous mutations that caused liver cancer growth with low prob-

ability [60]. However, another study demonstrated that chenodeox-

ycholic acid derivative HS-1200 could inhibit hepatocarcinogenesis in

rats and showed downregulation of MTH1 expression [61]. For MTH1

inhibitor development, we propose additional speciﬁc inhibitors from

our study, MI0639, MI1020 and MI1024, for further cancer research.

To a solution of NaOMe (10.09 g, 186.78 mmol) in absolute meth-

anol (42 mL) was added diethyl 2-ethylmalonate (5 mL, 26.67 mmol)

and guanidine hydrochloride (2.80 g, 29.34 mmol). The reaction

mixture was stirred at 85 ^◦C in sealed tube for 2 h. Methanol was

evaporated on a vacuum rotary evaporator and water (60 mL) was

added to the reaction mixture afterwards. The resulting colorless solu-

tion was subsequently neutralized by dropwise addition of acetic acid,

resulting in immediate and quantitative precipitation of the desired

product in the form of a ﬁne solid. The solid was ﬁltered off, washed

with water, ethanol, and acetone respectively. The product was dried

under high vacuum at 50 ^◦C for 12 h to yield 25 (3.14 g, 76%) as a white

powder. ¹H NMR (200 MHz, (CD₃)₂SO) δ 6.28 (s, 2H), 2.13 (q, J = 7.4

Hz, 2H), 0.87 (t, J = 7.4 Hz, 3H); ¹³C NMR (50 MHz, (CD₃)₂SO) δ 165.1

(2C), 153.1, 92.4, 16.3, 14.6; MS (m/z): [M - H]^-254.0.

4.1.2. 4,6-Dichloro-5-ethylpyrimidin-2-amine (4)

6-Chloro-5-ethyl-N⁴-methylpyrimidine-2,4-diamine (500 mg, 3.22

mmol) was suspended under inert atmosphere in a 2 M solution of the

Vilsmeier-Haack–Arnold reagent (25.78 mmol, 13 mL) in chloroform.

The reaction mixture was subsequently heated at reﬂux for 4 h. The

reaction mixture was cooled to room temperature, poured into ice and

rapidly neutralized with a saturated aqueous NaHCO₃solution. The

obtained mixture was partitioned between chloroform and water. The

layers were separated and the aqueous layer was extracted with chlo-

roform. The combined organic layer was washed with brine and dried

over MgSO₄. The MgSO₄was removed by ﬁltration and the ﬁltrate was

concentrated in vacuo. This crude residue was dissolved in the mixture

of 99% ethanol (7 mL) and 37% aqueous HCl (0.7 mL). The reaction

mixture was heated at 50 ^◦C for 2 h, during which a crystalline product

began to precipitate directly from the reaction mixture. After that, water

(10 mL) was added and the reaction mixture was intensively stirred for

10 min. The precipitated product was ﬁltered off and washed with a

water/ethanol mixture (1/1, 3 mL), 5% aqueous solution of NaHCO₃(3

mL) and a water/ethanol mixture (1/1, 3 mL), and dried under high

vacuum to yield compound 1 (428 mg, 70%) as a white solid. Mp 208 ^◦C;

¹H NMR (200 MHz, (CD₃)₂SO) δ 7.32 (s, 2H), 2.61 (q, J = 7.3 Hz, 2H),

1.06 (t, J = 7.3 Hz, 3H); ¹³C NMR (50 MHz, (CD₃)₂SO) δ 161.5, 161.4

(2C) , 119.6, 23.1, 13.6.

4. Experimental section

4.1. General Information for synthetic methods

In this work, all the reactions were processed under argon protection

with magnetic stirring. All the reagents used in the reactions were

bought from Acros, Aldrich, Alfa Aesar, Carbosynth, Fluka, TCI and

Matrix, and were used without prior puriﬁcation. All the solvents were

ACS grade. They were purchased from Merck, Mallinkrodt, or ECHO and

used without prior puriﬁcation. Anhydrous solvents were prepared

trough a SP-1 standalone solvent puriﬁcation system (LC Technology

Solutions) and contained less than 10 ppm water determined by Karl-

Fisher moisture analysis. During the reactions, the reacting progresses

were monitored by TLC, which was performed on Merck Kieselgel 60

F254 TLC plate. After the reactions ﬁnished, the products were puriﬁed

by extraction and then isolated by silica gel column chromatography

using Merck silica gel (40–60 μM). All the products were characterized

by ¹H NMR and ¹³C NMR spectra obtained from a Bruker spectrometer

(DPX-200). And the chemical shifts were referenced to the central peak

1

of DMSO‑d₆spectrum (2.49 ppm for H NMR and 39.52 ppm for ¹³C

NMR), CDCl₃spectrum (7.24 ppm for ¹H NMR and 77.16 ppm for ¹³

C

4.1.3. 6-Chloro-5-ethyl-N⁴-methylpyrimidine-2,4-diamine (5)

NMR), CD₃OD spectrum (3.31 ppm for ¹H NMR and 49.00 ppm for ¹³

NMR), or D₂O (4.79 ppm for ¹H NMR). Standard abbreviations used to

describe multiplicity of peaks in ¹H NMR were shown as follow: s =

singlet, d = doublet, t = triplet, q = quadruplet, m = multiplet and br =

broad. Melting points were measured by MEL-TEMP (Laboratory De-

vices Inc., USA) without calibration. Mass spectroscopy (MS) experi-

ments were performed on a Bruker spectrometer (Esquire2000) by

utilizing electrospray ionization and analyzed by quadrupole detector.

High resolution mass spectroscopy was performed on Bruker micro TOF

using ESI and time-of-ﬂight (TOF) detector. High-performance liquid

chromatograph (HPLC) analyses (Shimadzu LC-20AR) were obtained

To a solution of compound 1 (300 mg, 1.56 mmol) in isopropyl

alcohol (12 mL) was added methylamine hydrochloride (600 mg, 8.89

mmol) and DIPEA (1.3 mL). The reaction mixture was stirred at 100 ^◦C

for 17 h in sealed vial. After cooled to room temperature, the mixture

was reduced in volume under vacuum until most of the solvent had been

removed. The residue was partitioned between CH₂Cl₂and water. The

layers were separated and the aqueous layer was extracted with CH₂Cl₂.

The combined organic layer was washed with brine and dried over

MgSO₄. The MgSO₄was removed by ﬁltration and the ﬁltrate was

concentrated in vacuo. After that, the crude product was dissolved in

minimum amount of CH₂Cl₂and hexane was added to the solution

using a Phenomenex Kinetex XB-C18 column (5 μm, 100 × 4.6 mm) with

15

C. Peng et al.

Bioorganic Chemistry 110 (2021) 104813

dropwise until precipitation occurred. The mixture was left to settle

overnight, and the precipitate was collected by ﬁltration and dried

under high vacuum to afford 2 (182 mg, 62%) as an off-white solid. ¹H

NMR (200 MHz, CDCl₃) δ 4.75 (s, 3H), 2.95 (d, J = 4.9 Hz, 3H), 2.44 (q,

J = 7.6 Hz, 2H), 1.06 (t, J = 7.6 Hz, 3H); ¹³C NMR (50 MHz, CDCl₃) δ

162.9, 160.6, 157.3, 106.8, 28.9, 20.0, 12.5; MS (m/z): [M + H]⁺187.0.

CDCl₃) δ 166.1, 162.6, 160.8, 93.2, 23.7, 8.1 (2C); MS (m/z): [M + H]⁺

185.0.

4.1.8. 5-Ethyl-N⁴-methyl-6-morpholinopyrimidine-2,4-diamine (MI1001)

To a solution of compound 2 (100 mg, 0.54 mmol) in diphennyl ether

(3 mL) was added morpholine (0.1 mL, 1.07 mmol) and DIPEA (0.15 mL,

0.80 mmol). The reaction mixture was stirred at 210 ^◦C for 17 h in sealed

vial. The resulting mixture was puriﬁed by column chromatography

(silica gel: φ 2 × 8 cm; eluted by CH₂Cl₂/MeOH = 100/0 to 95/5). The

desired fraction (R_f= 0.31, CH₂Cl₂/MeOH = 95/5) was collected and

the solvent was evaporated to yield MI1001 (29 mg, 23%) as an orange

solid. Mp 129–132 ^◦C; ¹H NMR (200 MHz, CDCl₃) δ 4.51 (s, 2H), 4.39 (s,

1H), 3.76 (t, J = 4.8 Hz, 4H), 3.06 (t, J = 4.8 Hz, 4H), 2.95 (d, J = 4.8 Hz,

3H), 2.36 (q, J = 7.6 Hz, 2H), 1.10 (t, J = 7.6 Hz, 3H); ¹³C NMR (50 MHz,

CDCl₃) δ 166.4, 163.9, 160.9, 98.0, 67.6 (2C), 51.1 (2C), 28.8, 18.3,

13.2; HRMS m/z (ESI) calcd for C₁₁H₁₉N₅O [M + H]⁺238.1662; found,

238.1683. HPLC purity 97% (tR = 3.49 min).

4.1.4. 5-Ethyl-N⁴,N⁶-dimethylpyrimidine-2,4,6-triamine (6)

To a solution of compound 1 (130 mg, 0.68 mmol) in DMF (4.5 mL)

was added methylamine hydrochloride (70 mg, 1.02 mmol) and DIPEA

(0.4 mL, 2.03 mmol). The reaction mixture was stirred at 50 ^◦C for 28 h.

The mixture was reduced in volume under vacuum until most of the

solvent had been removed. The residue was partitioned between CH₂Cl₂

and water. The layers were separated and the aqueous layer was

extracted with CH₂Cl₂. The combined organic layer was washed with

brine and dried over MgSO₄. The MgSO₄was removed by ﬁltration and

the ﬁltrate was concentrated in vacuo. The resulting mixture was puri-

ﬁed by column chromatography (silica gel: φ 1.2 × 7 cm; eluted by

Hexane/EtOAc = 100/0 to 85/15). The desired fraction (R_f= 0.64,

Hexane/EtOAc = 50/50) was collected and the solvent was evaporated

to yield 31 (14 mg, 12%) as a white crystalline solid. Mp 153–154 ^◦C; ¹H

NMR (200 MHz, CDCl₃) δ 4.75 (s, 2H), 2.97 (s, 6H), 2.61 (q, J = 7.4 Hz,

2H), 1.17 (t, J = 7.4 Hz, 3H); ¹³C NMR (50 MHz, CDCl₃) δ 167.6, 161.3,

159.6, 110.1, 41.6 (2C), 22.2, 14.0.

4.1.9. 5-Ethyl-N⁴-methyl-6-(4-methylpiperazin-1-yl)pyrimidine-2,4-

diamine (MI1003)

To a solution of compound 2 (150 mg, 0.80 mmol) in diphennyl ether

(3 mL) was added 1-methylpiperazine (0.27 mL, 2.41 mmol) and DIPEA

(0.21 mL, 1.21 mmol). The reaction mixture was stirred at 240 ^◦C for 17

h in sealed vial. The resulting mixture was puriﬁed by column chro-

matography (silica gel: φ 2 × 8 cm; eluted by CH₂Cl₂/MeOH = 100/0 to

95/5). The desired fraction (R_f= 0.28, CH₂Cl₂/MeOH = 95/5) was

collected and the solvent was evaporated to yield MI1003 (97 mg, 48%)

as a pale yellow solid. Mp 162–165 ^◦C; ¹H NMR (200 MHz, CDCl₃) δ 4.48

(s, 2H), 4.34 (s, 1H), 3.12 (t, J = 5.3 Hz, 4H), 2.94 (d, J = 4.9 Hz, 3H),

2.48 (t, J = 5.3 Hz, 4H), 2.34 (q, J = 7.6 Hz, 2H), 2.30 (s, 3H), 1.10 (t, J

= 7.6 Hz, 3H); ¹³C NMR (50 MHz, CDCl₃) δ 166.5, 163.8, 160.8, 97.5,

55.8 (2C), 50.3 (2C), 46.7, 28.9, 18.5, 13.2; HRMS m/z (ESI) calcd for

C₁₂H₂₂N₆[M + H]⁺251.1979; found, 251.1998. HPLC purity 99% (tR

= 1.97 min).

4.1.5. 6-Chloro-N⁴-cyclopropyl-5-ethylpyrimidine-2,4-diamine (7)

To a solution of compound 1 (343 mg, 1.79 mmol) in n-butanol (10

mL) was added cyclopropylamine (0.37 mL, 5.36 mmol) and DIPEA

(0.78 mL, 4.46 mmol). The reaction mixture was stirred at 95 ^◦C for 18

h. After cooled to room temperature, the mixture was reduced in volume

under vacuum until most of the solvent had been removed. The residue

was poured into water, and the precipitate was collected by ﬁltration

and dried under high vacuum to give 32 (359 mg, 95%) as a yellow

solid. Mp 158–160 ^◦C; ¹H NMR (200 MHz, CDCl₃) δ 4.82 (s, 3H), 2.77

(m, 1H), 2.40 (q, J = 7.6 Hz, 2H), 1.03 (t, J = 7.6 Hz, 3H), 0.80 (m, 2H),

0.50 (m, 2H); ¹³C NMR (50 MHz, CDCl₃) δ 163.4, 160.6, 157.5, 106.8,

24.7, 20.0, 12.5, 8.0 (2C); MS (m/z): [M + H]⁺213.1.

4.1.10. 5-Ethyl-N⁴-methyl-6-(piperidin-1-yl)pyrimidine-2,4-diamine

(MI1006)

To a solution of compound 2 (130 mg, 0.70 mmol) in diphennyl ether

(3 mL) was added piperidine (0.35 mL, 3.54 mmol) and DIPEA (0.21 mL,

1.21 mmol). The reaction mixture was stirred at 240 ^◦C for 18 h in sealed

vial. The resulting mixture was puriﬁed by column chromatography

(silica gel: φ 2 × 8 cm; eluted by CH₂Cl₂/MeOH = 100/0 to 90/10). The

desired fraction (R_f= 0.18, CH₂Cl₂/MeOH = 95/5) was collected and

the solvent was evaporated to yield MI1006 (84 mg, 51%) as a pale

yellow solid. Mp 144–146 ^◦C; ¹H NMR (200 MHz, CDCl₃) δ 4.49 (s, 2H),

4.34 (s, 1H), 3.00 (m, 4H), 2.93 (d, J = 4.9 Hz, 3H) , 2.34 (q, J = 7.6 Hz,

2H), 1.57 (m, 6H), 1.09 (t, J = 7.6 Hz, 3H); ¹³C NMR (50 MHz, CDCl₃) δ

167.7, 163.7, 160.9, 98.0, 51.9 (2C), 28.8, 26.7 (2C), 25.1, 18.4, 13.2;

HRMS m/z (ESI) calcd for C₁₂H₂₁N₅[M + H]⁺236.1870; found,

236.1890. HPLC purity 99% (tR = 5.02 min).

4.1.6. 6-Chloro-N⁴-methylpyrimidine-2,4-diamine (9)

To a solution of 4,6-dichloropyrimidin-2-amine (1 g, 6.10 mmol) in

isopropyl alcohol (25 mL) was added methylamine hydrochloride (618

mg, 9.15) and DIPEA (3.2 mL, 10.98 mmol). The reaction mixture was

stirred at 100 ^◦C for 17 h in sealed vial. After cooled to room tempera-

ture, the mixture was reduced in volume under vacuum until most of the

solvent had been removed. The residue was partitioned between CH₂Cl₂

and water. The layers were separated and the aqueous layer was

extracted with CH₂Cl₂. The combined organic layer was washed with

brine and dried over MgSO₄. The MgSO₄was removed by ﬁltration and

the ﬁltrate was concentrated in vacuo to afford 6 (641 mg, 66%) as an

off-white solid. Mp 170 ^◦C; ¹H NMR (200 MHz, CDCl₃) δ 5.77 (s, 1H),

4.99 (s, 3H), 2.86 (d, J = 4.0 Hz, 3H); ¹³C NMR (50 MHz, (CD₃)₂SO) δ

165.4 (2C), 163.5, 93.7, 28.1; MS (m/z): [M + H]⁺159.0.

4.1.11. N⁴-Methyl-6-(4-methylpiperazin-1-yl)pyrimidine-2,4-diamine

(MI1012)

4.1.7. 6-Chloro-N⁴-cyclopropylpyrimidine-2,4-diamine (10)

To a solution of compound 6 (173 mg, 1.09 mmol) in diphennyl ether

(3 mL) was added 1-methylpiperazine (0.4 mL, 3.27 mmol) and DIPEA

(0.3 mL, 1.64 mmol). The reaction mixture was stirred at 150 ^◦C for 19

h. The resulting mixture was puriﬁed by column chromatography (silica

gel: φ 2 × 10 cm; eluted by CH₂Cl₂/MeOH = 100/0 to 95/5). The desired

fraction (R_f= 0.31, CH₂Cl₂/MeOH = 95/5) was collected and the sol-

vent was evaporated t₁o yield MI1012 (74 mg, 31%) as a pale yellow

solid. Mp 156–158 ^◦C; H NMR (200 MHz, CDCl₃) δ 4.96 (s, 1H), 4.44 (s,

3H), 3.52 (t, J = 5.2 Hz, 4H), 2.81 (d, J = 5.2 Hz, 3H), 2.42 (t, J = 5.3 Hz,

4H), 2.30 (s, 3H); ¹³C NMR (50 MHz, CDCl₃) δ 165.7, 164.8, 162.7, 72.9,

55.2 (2C), 46.6, 44.5 (2C), 28.9; HRMS m/z (ESI) calcd for C₁₀H₁₈N₆[M

+ H]⁺223.1666; found, 223.1688. HPLC purity 99% (tR = 3.43 min).

To a solution of 4,6-dichloropyrimidin-2-amine (1 g, 6.10 mmol) in

n-butanol (30 mL) was added cyclopropylamine (1.3 mL, 28.39 mmol)

and DIPEA (2.7 mL, 15.24 mmol). The reaction mixture was stirred at

95 ^◦C for 17 h. After cooled to room temperature, the mixture was

reduced in volume under vacuum until most of the solvent had been

removed. The residue was partitioned between CH₂Cl₂and water. The

layers were separated and the aqueous layer was extracted with CH₂Cl₂.

The combined organic layer was washed with brine and dried over

MgSO₄. The MgSO₄was removed by ﬁltration and the ﬁltrate was

concentrated in vacuo to give 7 (798 mg, 71%) as a yellow solid. Mp

180 ^◦C; ¹H NMR (200 MHz, CDCl₃) δ 6.11 (s, 1H), 5.42 (s, 1H), 5.01 (s,

2H), 2.47 (m, 1H), 0.80 (m, 2H), 0.56 (m, 2H); ¹³C NMR (50 MHz,

16

C. Peng et al.

Bioorganic Chemistry 110 (2021) 104813

4.1.12. N⁴-Cyclopropyl-5-ethyl-6-morpholinopyrimidine-2,4-diamine

(MI1013)

4.1.16. N⁴-Cyclopropyl-6-(4-methylpiperazin-1-yl)pyrimidine-2,4-diamine

(MI1022)

To a solution of compound 4 (120 mg, 0.56 mmol) in diphennyl ether

(3 mL) was added morpholine (0.15 mL, 1.69 mmol) and DIPEA (0.15

mL, 0.85 mmol). The reaction mixture was stirred at 210 ^◦C for 17 h in

sealed vial. The resulting mixture was puriﬁed by column chromatog-

raphy (silica gel: φ 2.5 × 8 cm; eluted by CH₂Cl₂/MeOH = 100/0 to 97/

3). The desired fraction (R_f= 0.37, CH₂Cl₂/MeOH = 95/5) was collected

and the solvent was evaporated to yield MI1013 (98 mg, 66%) as a pale

yellow solid. Mp 174–176 ^◦C; ¹H NMR (200 MHz, CDCl₃) δ 4.60 (s, 2H),

4.56 (s, 1H), 3.76 (t, J = 4.8 Hz, 4H), 3.06 (t, J = 4.8 Hz, 4H), 2.76 (m,

1H), 2.31 (q, J = 7.6 Hz, 2H), 1.05 (t, J = 7.6 Hz, 3H), 0.76 (m, 2H), 0.47

(m, 2H); ¹³C NMR (50 MHz, CDCl₃) δ 166.6, 164.2, 161.0, 97.9, 67.5

(2C), 51.0 (2C), 24.6, 18.3, 13.2, 8.1 (2C); HRMS m/z (ESI) calcd for

To a solution of compound 7 (190 mg, 1.03 mmol) in diphennyl ether

(3 mL) was added 1-methylpiperazine (0.4 mL, 3.09 mmol) and DIPEA

(0.27 mL, 1.55 mmol). The reaction mixture was stirred at 150 ^◦C for 23

h. The resulting mixture was puriﬁed by column chromatography (silica

gel: φ 2 × 10 cm; eluted by CH₂Cl₂/MeOH = 100/0 to 97/3). The desired

fraction (R_f= 0.31, CH₂Cl₂/MeOH = 95/5) was collected and the sol-

vent was evaporated to yield MI1022 (209 mg, 82%) as a pale yellow

solid. Mp 152–154 ^◦C; ¹H NMR (200 MHz, CDCl₃) δ 5.33 (s, 1H), 4.93 (s,

1H), 4.47 (s, 2H), 3.54 (t, J = 5.3 Hz, 4H), 2.42 (t, J = 5.3 Hz, 4H),2.42

(m, 1H), 0.70 (m, 2H), 0.52 (m, 2H); ¹³C NMR (50 MHz, CDCl₃) δ 165.9,

164.8, 162.7, 74.2, 55.3 (2C), 46.6 (2C), 44.55, 23.7, 7.9 (2C); HRMS m/

z (ESI) calcd for C₁₂H₂₀N₆[M + H]⁺249.1822; found, 249.1843. HPLC

purity 100% (tR = 8.21 min).

C

₁₃H₂₁N₅O [M + H]⁺264.1819; found, 264.1835. HPLC purity 99% (tR

= 7 min).

4.1.17. N⁴-Cyclopropyl-6-(piperidin-1-yl)pyrimidine-2,4-diamine

(MI1024)

4.1.13. N⁴-Cyclopropyl-5-ethyl-6-(4-methylpiperazin-1-yl)pyrimidine-2,4-

diamine (MI1014)

To a solution of compound 7 (200 mg, 1.08 mmol) in diphennyl ether

(3 mL) was added piperidine (0.32 mL, 3.25 mmol) and DIPEA (0.28 mL,

To a solution of compound 4 (120 mg, 0.56 mmol) in diphennyl ether

(3 mL) was added 1-methylpiperazine (0.25 mL, 2.26 mmol) and DIPEA

(0.15 mL, 0.86 mmol). The reaction mixture was stirred at 240 ^◦C for 17

h in sealed vial. The resulting mixture was puriﬁed by column chro-

matography (silica gel: φ 2.5 × 8 cm; eluted by CH₂Cl₂/MeOH = 100/

0 to 94/6). The desired fraction (R_f= 0.16, CH₂Cl₂/MeOH = 95/5) was

collected and the solvent was evaporated to yield MI1014 (77 mg, 50%)

as a yellow solid. Mp 126–130 ^◦C; ¹H NMR (200 MHz, CDCl₃) δ 4.67 (s,

2H), 4.53 (s, 1H), 3.10 (t, J = 5.2 Hz, 4H), 2.73 (m, 1H), 2.46 (t, J = 5.2

Hz, 4H), 2.28 (q, J = 7.6 Hz, 2H), 2.28 (s, 3H), 1.04 (t, J = 7.6 Hz, 3H),

0.75 (m, 2H), 0.45 (m, 2H); ¹³C NMR (50 MHz, CDCl₃) δ 166.6, 164.1,

160.9, 97.4, 55.7 (2C), 50.3 (2C), 46.6, 24.6, 18.4, 13.2, 8.1 (2C); HRMS

m/z (ESI) calcd for C₁₄H₂₄N₆[M + H]⁺277.2135; found, 277.2149.

HPLC purity 97% (tR = 3.49 min). HPLC purity 95% (tR = 7.92 min).

◦

1.62 mmol). The reaction mixture was stirred at 140 C for 17 h. The

resulting mixture was puriﬁed by column chromatography (silica gel: φ

2 × 8 cm; eluted by CH₂Cl₂/MeOH = 100/0 to 96/4). The desired

fraction (R_f= 0.26, CH₂Cl₂/MeOH = 95/5) was collected and the sol-

vent was evaporated to yield MI1024 (158 mg, 63%) as an orange foam.

¹H NMR (200 MHz, CDCl₃) δ 5.34 (s, 1H), 4.78 (s, 1H), 4.39 (s, 2H),3.50

(m, 4H), 2.44 (m, 1H), 1.60 (m, 6H), 0.72 (m, 2H), 0.53 (m, 2H); ¹³C

NMR (50 MHz, CDCl₃) δ 165.8, 164.7, 162.7, 74.0, 45.7 (2C), 26.0, 25.3

(2C), 23.7, 7.9 (2C); HRMS m/z (ESI) calcd for C₁₂H₁₉N₅[M + H]⁺

234.1713; found, 234.1734. HPLC purity 97% (tR = 12.19 min).

4.1.18. N⁴-Cyclopropyl-6-morpholinopyrimidine-2,4-diamine (MI1025)

To a solution of compound 7 (200 mg, 1.08 mmol) in diphennyl ether

(3 mL) was added morpholine (0.28 mL, 3.25 mmol) and DIPEA (0.3 mL,

4.1.14. N⁴-Cyclopropyl-5-ethyl-6-(piperidin-1-yl)pyrimidine-2,4-diamine

(MI1016)

1.62 mmol). The reaction mixture was stirred at 150 C for 13 h. The

◦

resulting mixture was puriﬁed by column chromatography (silica gel:

φ1.2 × 7.5 cm; eluted by CH₂Cl₂/MeOH = 100/0 to 96/4). The desired

fraction (R_f= 0.39, CH₂Cl₂/MeOH = 95/5) was collected and the sol-

vent was evaporated to yield MI1025 (163 mg, 64%) as a white solid.

Mp 164–170 ^◦C; ¹H NMR (200 MHz, CDCl₃) δ 5.31 (s, 1H), 5.17 (s, 1H),

4.66 (s, 2H), 3.74 (t, J = 5.3 Hz, 4H), 3.51 (t, J = 5.3 Hz, 4H),2.43 (m,

1H), 0.75 (m, 2H), 0.56 (m, 2H); ¹³C NMR (50 MHz, CDCl₃) δ 164.8,

164.7, 161.5, 73.8, 67.1 (2C), 45.0 (2C), 23.7, 7.9 (2C); HRMS m/z (ESI)

calcd for C₁₁H₁₇N₅O [M + H]⁺236.1506; found, 236.1525. HPLC purity

99% (tR = 10.1 min).

To a solution of compound 4 (98 mg, 0.46) in diphennyl ether (3 mL)

was added piperidine (0.19 mL, 1.88 mmol) and DIPEA (0.12 mL, 0.71

mmol). The reaction mixture was stirred at 240 ^◦C for 48 h in sealed vial.

The resulting mixture was puriﬁed by column chromatography (silica

gel: φ 1.2 × 7 cm; eluted by CH₂Cl₂/MeOH = 100/0 to 99/1). The

desired fraction (R_f= 0.54, CH₂Cl₂/MeOH = 95/5) was collected and

the solvent was evaporated to yield 33c (12 mg, 10%) as a yellow oil. ¹H

NMR (200 MHz, CDCl₃) δ 4.59 (s, 2H), 4.52 (s, 1H), 2.99 (t, J = 5.4 Hz,

4H), 2.75 (m, 1H), 2.29 (q, J = 7.6 Hz, 2H), 1.60 (m, 6H), 1.05 (t, J =

7.6 Hz, 3H), 0.75 (m, 2H), 0.45 (m, 2H); ¹³C NMR (50 MHz, CDCl₃) δ

167.7, 164.0, 160.9, 97.9, 51.8 (2C), 26.7 (2C), 25.1, 24.6, 18.5, 13.2,

8.1 (2C); MS (m/z): [M + H]⁺262.2. HPLC purity 99% (tR = 3.69 min).

4.1.19. N⁴-Methyl-6-(piperidin-1-yl)pyrimidine-2,4-diamine (MI1026)

To a solution of compound 6 (200 mg, 1.26 mmol) in diphennyl ether

(3 mL) was added piperidine (0.5 mL, 5.04 mmol) and DIPEA (0.33 mL,

4.1.15. N⁴-Methyl-6-morpholinopyrimidine-2,4-diamine (MI1020)

To a solution of compound 6 (105 mg, 0.66 mmol) in diphennyl ether

(3 mL) was added morpholine (0.17 mL, 1.99 mmol) and DIPEA (0.17

mL, 0.99 mmol). The reaction mixture was stirred at 135 ^◦C for 20 h. The

resulting mixture was puriﬁed by column chromatography (silica gel: φ

2 × 10 cm; eluted by CH₂Cl₂/MeOH = 100/0 to 98/2). The desired

fraction (R_f= 0.32, CH₂Cl₂/MeOH = 95/5) was collected and the sol-

vent was evaporated to yield MI1020 (83 mg, 60%) as a pale yellow

solid. Mp 158–160 ^◦C; ¹H NMR (200 MHz, CDCl₃) δ 4.92 (s, 1H), 4.78 (s,

1H), 4.63 (s, 2H), 3.72 (t, J = 5.2 Hz, 4H), 3.46 (t, J = 5.2 Hz, 4H), 2.80

(d, J = 4.0 Hz, 3H); ¹³C NMR (50 MHz, CDCl₃) δ 165.3, 165.0, 162.2,

72.9, 67.1 (2C), 45.0 (2C), 29.0; HRMS m/z (ESI) calcd for C₉H₁₅N₅O

[M + H]⁺210.1349; found, 210.1374. HPLC purity 99% (tR = 8.59

min).

1.89 mmol). The reaction mixture was stirred at 150 C for 17 h. the

◦

resulting mixture was puriﬁed by column chromatography (silica gel: φ

2 × 10 cm; eluted by CH₂Cl₂/MeOH = 100/0 to 96/4). The desired

fraction (R_f= 0.26, CH₂Cl₂/MeOH = 94/6) was collected and the sol-

vent was evaporated to yield MI1026 (131 mg, 50%) as a pale yellow

solid. Mp 154–156 ^◦C; ¹H NMR (200 MHz, CDCl₃) δ 4.95 (s, 1H), 4.47 (s,

3H), 3.47 (t, J = 5.7 Hz, 4H), 2.80 (d, J = 5.0 Hz, 3H), 1.58 (m, 6H); ¹³C

NMR (50 MHz, CDCl₃) δ 165.8, 164.7, 162.8, 72.8, 45.7 (2C), 29.0, 26.0

(2C), 25.3; HRMS m/z (ESI) calcd for C₁₀H₁₇N₅[M + H]⁺208.1557;

found, 208.1582. HPLC purity 98% (tR = 11.15 min).

4.1.20. N⁴-Cyclopropyl-6-(4-(oxetan-3-yl)piperazin-1-yl)pyrimidine-2,4-

diamine (MI1029)

To a solution of compound 7 (200 mg, 1.08 mmol) in diphennyl ether

(3 mL) was added 1-(oxetan-3-yl)piperazine (0.4 mL) and DIPEA (0.3

mL, 1.72 mmol). The reaction mixture was stirred at 150 ^◦C for 7 h. the

resulting mixture was puriﬁed by column chromatography (silica gel: φ

17

C. Peng et al.

Bioorganic Chemistry 110 (2021) 104813

2 × 10 cm; eluted by CH₂Cl₂/MeOH = 100/0 to 98/2). The desired

fraction (R_f= 0.40, CH₂Cl₂/MeOH = 95/5) was collected and the sol-

vent was evaporated to yield MI1029 (251 mg, 80%) as a white solid.

Mp 178–181 ^◦C; ¹H NMR (200 MHz, CDCl₃) δ 5.32 (s, 1H), 4.94 (s, 1H),

4.67 (t, J = 6.5 Hz, 2H), 4.62 (t, J = 6.5 Hz, 2H), 4.49 (s, 1H), 3.57 (t, J =

5.2 Hz, 4H), 3.48 (m, J = 6.3 Hz, 1H), 2.46 (m, 1H), 2.35 (t, J = 5.2 Hz,

4H), 0.71 (m, 2H), 0.51 (m, 2H); ¹³C NMR (50 MHz, CDCl₃) δ 166.0,

164.8, 162.7, 75.9 (2C), 74.2, 59.6, 49.8 (2C), 44.2 (2C), 23.7, 7.9 (2C);

HRMS m/z (ESI) calcd for C₁₄H₂₂N₆O [M + H]⁺291.1928; found,

291.1938. HPLC purity 99% (tR = 8.58 min).

days at 4 ^◦C. The protein mixture was further puriﬁed by size-exclusion

chromatography with a Superdex 75 (GE Healthcare) column in gel

ﬁltration buffer (20 mM Tris-HCl, pH 7.4, 150 mM NaCl, 5% (v/v)

glycerol, and 2 mM TCEP). The monomer form of His tag-free MTH1 was

collected and concentrated using a 10 kDa MWCO centrifugal concen-

trator (Merck Millipore) at 4 ^◦C. The concentrations of the puriﬁed

proteins were determined by measuring the absorbance at 280 nm

(DeNovix DS-11) and calculated using the extinction coefﬁcient 28,120

M^{ꢀ 1}cm^{ꢀ 1}, and the proteins samples were stored for crystallization and

biochemical analysis.

4.1.21. N⁴-Methyl-6-(4-(oxetan-3-yl)piperazin-1-yl)pyrimidine-2,4-

diamine (MI1030)

4.3. Identiﬁcation of MTH1 inhibitor hits by uHTS of a large compound

library

To a solution of compound 6 (200 mg, 1.26 mmol) in diphennyl ether

(3.5 mL) was added 1-(oxetan-3-yl)piperazine (0.4 mL) and DIPEA (0.3

mL, 1.89 mmol). The reaction mixture was stirred at 140 ^◦C for 16 h. the

resulting mixture was puriﬁed by column chromatography (silica gel: φ

2 × 10 cm; eluted by CH₂Cl₂/MeOH = 100/0 to 96/4). The desired

fraction (R_f= 0.25, CH₂Cl₂/MeOH = 95/5) was collected and the sol-

vent was evaporated to yield MI1030 (250 mg, 75%) as a white solid.

Mp 180–182 ^◦C; ¹H NMR (200 MHz, CDCl₃) δ 4.95 (s, 1H), 4.67 (t, J =

6.5 Hz, 2H), 4.62 (t, J = 6.5 Hz, 2H), 4.45 (s, 3H), 3.55 (t, J = 5.2 Hz,

4H), 3.48 (m, J = 6.6 Hz, 1H), 2.81 (d, J = 5.2 Hz, 3H), 2.34 (t, J = 5.2

Hz, 4H) ; ¹³C NMR (50 MHz, CDCl₃) δ 165.8, 164.8, 162.7, 75.9 (2C),

73.1, 59.6, 49.8 (2C), 44.2 (2C), 29.0; HRMS m/z (ESI) calcd for

C₁₂H₂₀N₆O [M + H]⁺265.1771; found, 265.1786. HPLC purity 95% (tR

= 2.36 min).

A fragment-based library with 2,313 compounds was ordered from

ChemBridge Corporation. The compounds were ﬁrst solubilized in 100%

dimethylsulfoxide (DMSO) and then adjusted to 2 mM for screening. The

screening was carried out with 1,536-well plates using the uHTS system

(GNF, San Diego, CA, USA) at the Genomics Research Center of

Academia Sinica in Taiwan. For MTH1 inhibition screening, the reaction

conditions were a total volume of 5

with 12.5

μL containing 0.45 nM MTH1 mixed

μ

M 8-oxodGTP (TriLink BioTechnologies) and 2 mM of the

compounds from the library in MTH1 reaction buffer (10 mM Tris-

acetate, pH 7.5, 40 mM NaCl, 10 mM Mg-acetate, 1 mM dithiothreitol

(DTT), and 0.005% (v/v) Tween 20) at 37 ^◦C for 75 min. After the re-

action time, 2.5 μL of converting reagent from the PPiLight inorganic

pyrophosphate assay kit (Lonza) was added to the sample mixture,

which was then allowed to incubate at 25 ^◦C for 30 min. Afterward, 2.5

μ

L of detecting reagent from the kit was added to the mixture followed

4.2. Cloning, expression, and puriﬁcation of MTH1

by incubation at 25 ^◦C for 30 min. The ﬁnal reaction volume was 10

μL

in each well of a 1536-well white plate. The luminescence was measured

by a ViewLux plate reader (Perkin Elmer Inc.). The integrated reading

time was set to 0.1 sec. The IC₅₀values were determined using nonlinear

regression and Prism software (GraphPad).

The coding sequence (ATGGGTGCCAGCCGCCTGTATACCCTGGT

TCTGGTGCTGCAGCCGCAGCGTGTTCTGCTGGGCATGAAAAAACGCGG

CTTTGGTGCCGGTCGTTGGAATGGTTTTGGCGGCAAAGTGCAAGAAGG

TGAAACCATTGAAGATGGCGCCCGCCGTGAACTGCAGGAAGAAAGCG

GTCTGACCGTGGATGCCCTGCATAAAGTGGGCCAGATCGTGTTTGAGT

TTGTGGGCGAACCGGAACTGATGGACGTGCATGTGTTTTGCACCGACA

GCATTCAGGGCACCCCGGTGGAAAGCGATGAAATGCGTCCGTGCTGGT

TCCAGCTGGATCAGATCCCGTTTAAGGACATGTGGCCGGATGATAGCT

ATTGGTTCCCGCTGCTGCTGCAGAAGAAAAAGTTCCACGGCTACTTCA

AATTCCAGGGCCAGGATACCATCCTGGACTATACTCTGCGCGAAGTGG

ATACCGTGTAG) of full-length human MTH1 (NUDT1) in pUC57 vector

was synthesized by Genewiz. MTH1 in-frame was ampliﬁed from the

pUC57-MTH1 construct by polymerase chain reaction and then subcl-

oned into the bacterial expression vector pET28a(+) (Novagen), in

which 6-His tag and thrombin cleavage site are included at the N-ter-

minus of MTH1. This plasmid pET28a-MTH1 was transformed into

E. coli (Rosetta) cells. For expression, cultures were grown in LB medium

4.4. PPiLight biochemical assay

MTH1 activity assay conditions were 0.45 nM MTH1 mixed with

12.5 μM substrate 8-oxo-dGTP (Trilink Biotechnologies) in MTH1 re-

action buffer (10 mM Tris-acetate, pH 7.5, 40 mM NaCl, 10 mM Mg-

acetate, 1 mM DTT, and 0.005% (v/v) Tween 20) at 37 ^◦C for 75 min.

The reaction volume was 40 μL. After the reaction time was ﬁnished, 20

μ

L of converting reagent from the PPiLight inorganic pyrophosphate

assay kit (Lonza) was added to the mixture at 25 ^◦C for 30 min, and then

another 20 L of detecting reagent from the kit was added and the

mixture was incubated at 25 ^◦C for 30 min. The total ﬁnal reaction

volume was 80 L. Detection was performed using a 96-well white plate,

μ

and luminescence was measured with an Inﬁnite200® PRO reader

(TECAN). The integrated reading time was set to 0.1 sec per sample. For

the inhibition assay, dose–response curves were recorded using con-

centrations ranging from 2 mM to 0.1 nM in 1/4 serial dilution steps for

MI0639, MI0320, and MI0652 and with concentrations ranging from 2

with kanamycin (50 μg/mL) and chloramphenicol (34 μg/mL). When

bacterial cell growth reached OD₆₀₀0.4–0.5 at 37 ^◦C, isopropyl β-D-1-

thiogalactopyranoside was added to a ﬁnal concentration of 1 mM, and

protein expression was induced at 22 ^◦C for 16 hr. Cells were harvested

by centrifugation at 7750g and stored at ꢀ 80 ^◦C until further use. The

E. coli cell pellet was resuspended in lysis buffer (20 mM Tris-HCl, pH

7.4, 500 mM NaCl, 10% (v/v) glycerol, 10 mM imidazole, 2 mM Tris(2-

carboxyethyl)phosphine (TCEP), and 2 mM phenylmethylsulfonyl

ﬂuoride). Cells were lysed by sonication (Misonix sonicator XL), and the

cell debris was removed via centrifugation at 18,900g for 25 min at 4 ^◦C

mM to 1 μM in 1/2 serial dilution steps for MI0861. A reaction with 8.4

μ

M TH588 served as a positive control for inhibition. The IC₅₀and Ki

values were determined by nonlinear regression using Prism software

(GraphPad).

(Beckman J2-MC). The supernatant was ﬁltered using a 0.45

μm ﬁlter,

4.5. Crystallization

and 5 mL was applied to a HisTrapFF column (GE Healthcare). The

column was washed to baseline, and the protein was eluted in a gradient

with elution buffer (lysis buffer plus 500 mM imidazole). The eluted

fractions were concentrated using Amicon® Ultra (MWCO 10 kDa;

Merck Millipore) centrifugal concentrator (Thermo Legend Mach 1.6R)

at 4 ^◦C, and the solution buffer was replaced with storage buffer (20 mM

HEPES pH 7.4, 300 mM NaCl, 10% (v/v) glycerol, 1 mM TCEP). Sub-

MTH1 protein at a concentration of 7 mg/mL in gel ﬁltration buffer

(20 mM Tris–HCl, pH 7.4, 150 mM NaCl, 5% (v/v) glycerol, 2 mM TCEP,

and 5 mM MgCl₂) was crystallized through the hanging-drop method at

20 ^◦C using a reservoir solution (30% PEG6000, 0.1 M sodium acetate,

pH 3.75, and 0.2 M lithium sulfate). In each 0.5 μL hanging drop, the

ratio of reservoir to protein in solution was 1:1 or 2:1. The volume of

mother liquid in the well was 1 mL. Long rectangular crystals of

sequently, 0.15 μg/μL thrombin was added to cut off the His tag for 2

18

C. Peng et al.

Bioorganic Chemistry 110 (2021) 104813

diffraction quality appeared about 3 days after crystalization. For

compound MI0639, MTH1 protein solution was ﬁrst mixed with 4 mM

compound for 2 hr at 25 ^◦C, and then the co-crystallization method was

used to grow crystals. For other compounds, MTH1 crystals were

transferred to a soaking solution containing 2 mM compounds in a buffer

with 30% (w/v) PEG6000, 200 mM lithium sulphate, 100 mM sodium

acetate at pH 3.75, and 10% (v/v) dimethylsulfoxide at 25 ^◦C. For the

compounds with lower micromolar IC₅₀values, compounds were used at

10 mM in the soaking experiment. After 18 hr of incubation, crystals

were removed from the soaking solution and ﬂash frozen directly in

liquid nitrogen.

(C), and cell growth in the presence of compound (Tx), the percentage

growth inhibition was calculated as [1 ꢀ (Tx ꢀ T0)/(C ꢀ T0)] × 100%

[70].

4.9. Flow cytometric assay of propidium iodide (PI) staining

Cells with the indicated treatments were harvested by trypsinization,

◦

ﬁxed with 70% (v/v) alcohol at ꢀ 20 C for 30 min and washed with

phosphate-buffered saline (PBS). After centrifugation, cells were re-

suspended with 0.5 mL PI solution containing Triton X-100 (0.1% v/

v), RNase (100 μg/mL) and PI (80 μg/mL). DNA content was analyzed

with FACScan and CellQuest software (Becton Dickinson, Mountain

View, CA). Ten thousand events were collected per sample and the data

were analyzed using ModFit LT version 3.3 [70].

4.6. Data collection, structure determination, and reﬁnement

The protein crystal diffraction data were collected at beamlines

15A1, 13B1, and 13C1 of the Taiwan Light Source at the National

Synchrotron Radiation Research Center (Hsinchu, Taiwan), collected

with a Rayonix MX-300 CCD detector, and compiled using Blu-Ice

software [62]. All diffraction data were processed for indexing, inte-

gration, and scaling in sequence using the HKL2000 program [63]. For

structure determination, the molecular replacement method was per-

formed using Phenix MR phaser software [64] using as the search model

the previously solved apoMTH1 structure with elimined water mole-

cules (Protein Data Bank code 3ZR1). Compound structures and geom-

etry restraints were generated using the PRODRG server [65], and the

compound structures were positioned in the MTH1 binding pocket using

LigandFit in the Phenix environment [66]. Model rebuilding and

structure reﬁnement were performed in Coot [67] and Phenix [68]. All

structural models were generated in PyMOL [69]. Detailed diffraction

and reﬁnement statistics are listed in Table 1. The electrostatic surfaces

of the structures were calculated using APBS software [44,45] and the

binding pocket volume was computed by the CASTp server [46].

4.10. Statistical analysis of MTH1 expression and survival analysis

The relationships between MTH1 mRNA expression in cancer cells

and adjacent cells were determined using the paired Student t test.

Survival curves of high and low MTH1 mRNA expression were plotted

using the Kaplan–Meier method and compared using the log-rank test.

The data were obtained from the Oncomine database. Data are pre-

sented as mean ± SD. A value of p < 0.05 or less was considered to be

statistically signiﬁcant [71].

5. Accession codes

Accession codes in the RCSB Protein Data Bank (PDB) for human

MTH1 with bound inhibitors: apoMTH1 (6JVF) and compounds MI0639

(6JVG), MI0320 (6JVH), MI0861 (6JVI), MI1006 (6JVJ), MI1012

(6JVK), MI1014 (6JVL), MI1016 (6JVM), MI1020 (6JVN), MI1022

(6JVO), MI1024 (6JVP), MI1025 (6JVQ), MI1026 (6JVR), MI1029

(6JVS), and MI1030 (6JVT).

4.7. Determination of binding constants by ITC

Author contributions

Prior to the ITC experiment, both MTH1 and compounds were

transferred into the ITC reaction buffer (100 mM Tris-acetate, pH 7.9,

40 mM NaCl, 10 mM MgCl₂, 0.005% Tween-20, and 1 mM DTT). The

The manuscript was written with contributions from all authors. All

authors have given approval to the ﬁnal version of the manuscript.

sample cell was ﬁlled with 60

saturation with 600 M to 860

μM MTH1 to 280

μL and then titrated to

μ

μM of compound solution using a syringe.

Declaration of Competing Interest

Each reaction was performed on a MicroCal-iTC200 instrument (GE

Healthcare) with 30 injections of 1.3 µL compound solution at 3 min

intervals followed by incubation at 25 ^◦C with stirring at 1000 rpm. The

data was analyzed by Origin 7.0 (OriginLab™), and curve ﬁtting was

performed using one-site ﬁtting mode. The values of ΔG, ΔH, ꢀ TΔS, and

K_dcould be calculated and determined. All values were collected by

three independent expriments. LE values of the data were calculated on

the basis of ΔG of a ligand for the number of HA, calculated as LE = ΔG/

HA [48].

The authors declare that they have no known competing ﬁnancial

interests or personal relationships that could have appeared to inﬂuence

the work reported in this paper.

Acknowledgements

High-throughput screening was carried out at the Genomics

Research Center of Academia Sinica in Taiwan. X-ray diffraction was

carried out at beam lines 15A1, 13B1, and 13C1 of the National Syn-

chrotron Radiation Research Center (Taiwan). Financial support was

received by Yi-Sheng Cheng from the Taiwan Ministry of Science and

Technology (MOST 109-2311-B-002-012 and MOST 109-2311-B-002-

026) and from the National Taiwan University (NTUCC-107L893105).

The authors thank Dr. Jih-Hwa Guh from Department of Pharmacy,

National Taiwan University for performing cancer cell survival assay.

4.8. Inhibition of cell proliferation

Human hormone-refractory prostate cancer cell lines PC-3 and

DU145 were obtained from the American Type Culture Collection

(Rockville, MD). Cells were cultured in RPMI-1640 medium with 10%

fetal bovine serum (v/v), penicillin (100 units/mL), and streptomycin

(100 μg/mL). Cultures were maintained in a humidiﬁed incubator at

37 ^◦C in 5% CO₂and 95% air. For inhibition of cell proliferation, cells

were seeded into 96-well plates. After 24 hr, cells were ﬁxed with 10%

(v/v) trichloroacetic acid (TCA) to represent the cell population at time

zero (T0). After additional incubation with 0.1% (v/v) DMSO or the

compound for 48 hr or 72 hr, depending on the cell doubling time, cells

were ﬁxed with 10% (v/v) TCA, and 0.4% (w/v) sulforhodamine B

(SRB) in 1% acetic acid was added to stain the cells. Unbound SRB was

washed out. SRB-bound cells were solubilized with 10 mM Trizma base.

Using the absorbance (515 nm) measurements for T0, control growth

Appendix A. Supplementary material

Supplementary data to this article can be found online at https://doi.

org/10.1016/j.bioorg.2021.104813.

References

[1] C. Gorrini, I.S. Harris, T.W. Mak, Modulation of oxidative stress as an anticancer

strategy, Nat. Rev. Drug Discov. 12 (2013) 931 –947 .

19

C. Peng et al.

Bioorganic Chemistry 110 (2021) 104813

[2] G.Y. Liou, P. Storz, Reactive oxygen species in cancer, Free Radi. Res. 44 (2010)

479 –496 .

[26] M. Carter, A.-S. Jemth, A. Hagenkort, B.D.G. Page, R. Gustafsson, J.J. Griese,

¨

H. Gad, N.C.K. Valerie, M. Desroses, J. Bostrom, U. Warpman Berglund,

[3] J.E. Klaunig, L.M. Kamendulis, B.A. Hocevar, Oxidative stress and oxidative

damage in carcinogenesis, Toxicol. Pathol. 38 (2010) 96 –109 .

[4] K.D. Carlson, M.T. Washington, Mechanism of efﬁcient and accurate nucleotide

incorporation opposite 7, 8-dihydro-8-oxoguanine by Saccharomyces cerevisiae

T. Helleday, P. Stenmark, Crystal structure, biochemical and cellular activities

demonstrate separate functions of MTH1 and MTH2, Nat. Commun. 6 (1) (2015),

https://doi.org/10.1038/ncomms8871.

[27] G. Samaranayake, M. Huynh, P. Rai, MTH1 as a Chemotherapeutic Target: The

Elephant in the Room, Cancers 9 (12) (2017) 47, https://doi.org/10.3390/

cancers9050047.

DNA polymerase η , Mol. Cell. Biol. 25 (6) (2005) 2169 –2176 .

[5] T. Hirano, K. Tamae, Differentiation of embryonic stem cells and oxidative DNA

damage, DNA repair systems, J. Stem. Cell Res. Ther. S 10 (2012) 1 –5 .

[6] K. Fujikawa, H. Kamiya, H. Yakushiji, Y. Nakabeppu, H. Kasai, Human MTH1

protein hydrolyzes the oxidized ribonucleotide, 2-hydroxy-ATP, Nucleic Acids Res.

29 (2001) 449 –454 .

¨

[28] L.M. Svensson, A.-S. Jemth, M. Desroses, O. Loseva, T. Helleday, M. Hogbom,

P. Stenmark, Crystal structure of human MTH1 and the 8-oxo-dGMP product

complex, FEBS Lett. 585 (2011) 2617 –2621 .

[29] X. Zhang, W. Song, Y. Zhou, F. Mao, Y. Lin, J. Guan, Q. Sun, Expression and

function of MutT homolog 1 in distinct subtypes of breast cancer, Oncol. Lett. 13

(2017) 2161 –2168 .

[7] K. Sakumi, M. Furuichi, T. Tsuzuki, T. Kakuma, S. Kawabata, H. Maki,

M. Sekiguchi, Cloning and expression of cDNA for a human enzyme that hydrolyzes

8-oxo-dGTP, a mutagenic substrate for DNA synthesis, J. Biol. Chem. 268 (1993)

23524 –23530 .

[30] Z. Versano, E. Shany, S. Freedman, L. Tuval-Kochen, M. Leitner, S. Paglin,

A. Toren, M. Yalon, MutT homolog 1 counteracts the effect of anti-neoplastic

treatments in adult and pediatric glioblastoma cells, Oncotarget 9 (44) (2018)

27547 –27563 .

[8] M. Furuichi, M.C. Yoshida, H. Oda, T. Tajiri, Y. Nakabeppu, T. Tsuzuki,

M. Sekiguchi, Genomic structure and chromosome location of the human mutT

homologue gene MTH1 encoding 8-oxo-dGTPase for prevention of A: T to C: G

transversion, Genomics 24 (1994) 485 –490 .

[31] X.-L. Shi, Y. Li, L.-M. Zhao, L.-W. Su, G. Ding, Delivery of MTH1 inhibitor (TH287)

and MDR1 siRNA via hyaluronic acid-based mesoporous silica nanoparticles for

oral cancers treatment, Colloids Surf. B: Biointerfaces 173 (2019) 599 –606 .

[32] T. Kawamura, M. Kawatani, M. Muroi, Y. Kondoh, Y. Futamura, H. Aono,

M. Tanaka, K. Honda, H. Osada, Proteomic proﬁling of small-molecule inhibitors

reveals dispensability of MTH1 for cancer cell survival, Sci. Rep. 6 (2016) 26521.

[33] H. Gad, O. Mortusewicz, S.G. Rudd, A. Stolz, N. Amaral, L. Brautigham, L. Pudelko,

[9] M.J. Bessman, D.N. Frick, S.F. O ’Handley, The MutT proteins or “Nudix ”

hydrolases, a family of versatile, widely distributed, “housecleaning ” enzymes,

J. Biol. Chem. 271 (1996) 25059 –25062 .

[10] D. Kang, J.-I. Nishida, A. Iyama, Y. Nakabeppu, M. Furuichi, T. Fujiwara,

M. Sekiguchi, K. Takeshige, Intracellular localization of 8-oxo-dGTPase in human

cells, with special reference to the role of the enzyme in mitochondria, J. Biol.

Chem. 270 (1995) 14659 –14665 .

´

¨

K. Sanjiv, C. Kalderen, A.-S. Jemth, I. Almlof, T. Visnes, N. Schultz, J. Bostrom, J.

M. Calderon Montano, A. Hagenkort, P. Groth, O. Loseva, C. Gokturk,

¨

[11] Y. Sakai, M. Furuichi, M. Takahashi, M. Mishima, S. Iwai, M. Shirakawa,

Y. Nakabeppu, A molecular basis for the selective recognition of 2-hydroxy-dATP

and 8-oxo-dGTP by human MTH1, J. Biol. Chem. 277 (10) (2002) 8579 –8587 .

[12] E. Coskun, P. Jaruga, A.-S. Jemth, O. Loseva, L.D. Scanlan, A. Tona, M.

S. Lowenthal, T. Helleday, M. Dizdaroglu, Addiction to MTH1 protein results in

intense expression in human breast cancer tissue as measured by liquid

chromatography-isotope-dilution tandem mass spectrometry, DNA Repair 33

(2015) 101 –110 .

T. Koolmeister, P. Wakchaure, E. Homan, C.E. Strom, M. Scobie, H. Bastians, U.

W. Berglund, T. Helleday, MTH1 promotes mitotic progression to avoid oxidative

DNA damage in cancer cells, BioRxiv (2019) 575290 .

´

[34] U. Warpman Berglund, K. Sanjiv, H. Gad, C. Kalderen, T. Koolmeister, T. Pham,

C. Gokturk, R. Jafari, G. Maddalo, B. Seashore-Ludlow, A. Chernobrovkin,

¨

A. Manoilov, I.S. Pateras, A. Rasti, A.-S. Jemth, I. Almlof, O. Loseva, T. Visnes, B.

O. Einarsdottir, F.Z. Gaugaz, A. Saleh, B. Platzack, O.A. Wallner, K.S.A. Vallin,

M. Henriksson, P. Wakchaure, S. Borhade, P. Herr, Y. Kallberg, P. Baranczewski, E.

´

¨

[13] S. Borrego, A. Vazquez, F. Dasí, C. Cerda, A. Iradi, C. Tormos, J. Sanchez, L. Bagan,

J. Homan, E. Wiita, V. Nagpal, T. Meijer, N. Schipper, S.G. Rudd, L. Brautigam,

´

J. Boix, C. Zaragoza, J. Camps, G. Saez, Oxidative Stress and DNA Damage in

A. Lindqvist, A. Filppula, T.-C. Lee, P. Artursson, J.A. Nilsson, V.G. Gorgoulis,

¨

Human Gastric Carcinoma: 8-Oxo-7 ’8-dihydro-2 ’-deoxyguanosine (8-oxo-dG) as a

Possible Tumor Marker, Int. J. Mol. Sci. 14 (2) (2013) 3467 –3486 .

[14] W.-J. Song, P. Jiang, J.-P. Cai, Z.-Q. Zheng, Expression of Cytoplasmic 8-oxo-Gsn

and MTH1 Correlates with Pathological Grading in Human Gastric Cancer, Asian

Pac. J. Cancer Prev. 16 (15) (2015) 6335 –6338 .

J. Lehtio, R.A. Zubarev, M. Scobie, T. Helleday, Seashore-Ludlow, Validation and

development of MTH1 inhibitors for treatment of cancer, Ann. Oncol. 27 (12)

(2016) 2275 –2283 .

[35] B.O. Einarsdottir, J. Karlsson, E.M.V. Soderberg, M.F. Lindberg, E. Funck-Brentano,

H. Jespersen, S.F. Brynjolfsson, R.O. Bagge, L. Carstam, M. Scobie, T. Koolmeister,

O. Wallner, U. Stierner, U.W. Berglund, L. Ny, L.M. Nilsson, E. Larsson, T. Helleday,

J.A. Nilsson, A patient-derived xenograft pre-clinical trial reveals treatment

responses and a resistance mechanism to karonudib in metastatic melanoma, Cell

Death Dis. 9 (2018) 810 .

[15] S. Akiyama, H. Saeki, Y. Nakashima, M. Iimori, H. Kitao, E. Oki, Y. Oda,

Y. Nakabeppu, Y. Kakeji, Y. Maehara, Prognostic impact of MutT homolog-1

expression on esophageal squamous cell carcinoma, Cancer Med. 6 (1) (2017)

258 –266 .

[16] Y. Tu, Z. Wang, X. Wang, H. Yang, P. Zhang, M. Johnson, N. Liu, H. Liu, W. Jin,

Y. Zhang, Birth of MTH1 as a therapeutic target for glioblastoma: MTH1 is

indispensable for gliomatumorigenesis, Am. J. Transl. Res. 8 (2016) 2803 .

[17] Y. Kumagae, M. Hirahashi, K. Takizawa, H. Yamamoto, M. Gushima, M. Esaki,

T. Matsumoto, M. Nakamura, T. Kitazono, Y. Oda, Overexpression of MTH1 and

OGG1 proteins in ulcerative colitis-associated carcinogenesis, Oncol. Lett. 16

(2018) 1765 –1776 .

[36] K.V. Huber, E. Salah, B. Radic, M. Gridling, J.M. Elkins, A. Stukalov, A.-S. Jemth,

¨

C. Gokturk, K. Sanjiv, K. Stromberg, Stereospeciﬁc targeting of MTH1 by (S)-

crizotinib as anticancer strategy, Nature 508 (2014) 222 .

[37] X. Qing, Z. Shao, X. Lv, F. Pu, F. Gao, L. Liu, D. Shi, Anticancer effect of (S)-

crizotinib on osteosarcoma cells by targeting MTH1 and activating reactive oxygen

species, Anti-cancer Drugs 29 (2018) 341 –352 .

[38] J.G. Kettle, H. Alwan, M. Bista, J. Breed, N.L. Davies, K. Eckersley, S. Fillery, K.

M. Foote, L. Goodwin, D.R. Jones, Potent and selective inhibitors of MTH1 probe

its role in cancer cell survival, J. Med. Chem. 59 (2016) 2346 –2361 .

[39] F. Rahm, J. Viklund, L. Tresaugues, M. Ellermann, A. Giese, U. Ericsson,

R. Forsblom, T. Ginman, J. Gunther, K. Hallberg, J. Lindstrom, L.B. Persson,

C. Silvander, A. Talagas, L. Diaz-Saez, O. Fedorov, K.V.M. Huber, I. Panagakou,

P. Siejka, M. Gorjanacz, M. Bauser, M. Andersson, Creation of a Novel Class of

Potent and Selective MutT Homologue 1 (MTH1) Inhibitors Using Fragment-Based

Screening and Structure-Based Drug Design, J. Med. Chem. 61 (2018) 2533 –2551 .

[40] T. Yokoyama, R. Kitakami, M. Mizuguchi, Discovery of a new class of MTH1

inhibitor by X-ray crystallographic screening, Eur. J. Med. Chem. 167 (2019)

153 –160 .

[18] J. Xu, Z.Y. Yang, X. Chen, X. Liu, HMTH1 induces the metastasis and recurrence of

the parotid adenoma by repairing DNA damage, Eur. Rev. Med. Pharmacol. Sci. 22

(2018) 4363 –4370 .

[19] A. Patel, D.G.A. Burton, K. Halvorsen, W. Balkan, T. Reiner, C. Perez-Stable,

A. Cohen, A. Munoz, M.G. Giribaldi, S. Singh, D.J. Robbins, D.M. Nguyen, P. Rai,

MutT Homolog 1 (MTH1) maintains multiple KRAS-driven pro-malignant

pathways, Oncogene 34 (20) (2015) 2586 –2596 .

[20] P. Rai, T.T. Onder, J.J. Young, J.L. McFaline, B. Pang, P.C. Dedon, R.A. Weinberg,

Continuous elimination of oxidized nucleotides is necessary to prevent rapid onset

of cellular senescence, Proc. Natl. Acad. Sci. U.S.A. 106 (1) (2009) 169 –174 .

[21] P. Rai, J.J. Young, D.G.A. Burton, M.G. Giribaldi, T.T. Onder, R.A. Weinberg,

Enhanced elimination of oxidized guanine nucleotides inhibits oncogenic RAS-

induced DNA damage and premature senescence, Oncogene 30 (12) (2011)

[41] P. Jansa, A. Holý, M. Dracínský, V. Kolman, Z. Janeba, P. Kostecka, E. Kmoníckova,

Z. Zídek, 5-Substituted 2-amino-4, 6-dihydroxypyrimidines and 2-amino-4, 6-

dichloropyrimidines: synthesis and inhibitory effects on immune-activated nitric

oxide production, Med. Chem. Res. 23 (2014) 4482 –4490 .

[22] M.G. Giribaldi, A. Munoz, K. Halvorsen, A. Patel, P. Rai, MTH1 expression is

required for effective transformation by oncogenic HRAS, Oncotarget 6 (13) (2015)

11519 –11529 .

[42] R.J. Altenbach, R.M. Adair, B.M. Bettencourt, L.A. Black, S.R. Fix-Stenzel, S.

M. Gopalakrishnan, G.C. Hsieh, H. Liu, K.C. Marsh, M.J. McPherson, I. Milicic, T.

R. Miller, T.A. Vortherms, U. Warrior, J.M. Wetter, N. Wishart, D.G. Witte,

P. Honore, T.A. Esbenshade, A.A. Hancock, J.D. Brioni, M.D. Cowart,

Structure ꢀ Activity Studies on a Series of a 2-Aminopyrimidine-Containing

Histamine H4 Receptor Ligands, J. Med. Chem. 51 (2008) 6571 –6580 .

[43] M.M. El-Gendy, M. Shaaban, K.A. Shaaban, A.M. El-Bondkly, H. Laatsch,

Essramycin: a ﬁrst triazolopyrimidine antibiotic isolated from nature, J. Antibiot.

61 (2008) 149 .

[23] P. Rai, Human Mut T Homolog 1 (MTH1): a roadblock for the tumor-suppressive

effects of oncogenic RAS-induced ROS, Small GTPases 3 (2) (2012) 120 –125 .

[24] T. Tsuzuki, A. Egashira, S. Kura, Analysis of MTH1 gene function in mice with

targeted mutagenesis, Mutat. Res. 477 (1-2) (2001) 71 –78 .

¨

[25] H. Gad, T. Koolmeister, A.-S. Jemth, S. Eshtad, S.A. Jacques, C.E. Strom, L.

¨

M. Svensson, N. Schultz, T. Lundback, B.O. Einarsdottir, A. Saleh, C. Goktürk,

¨

P. Baranczewski, R. Svensson, R.-A. Berntsson, R. Gustafsson, K. Stromberg,

K. Sanjiv, M.-C. Jacques-Cordonnier, M. Desroses, A.-L. Gustavsson, R. Olofsson,

[44] S. Unni, Y. Huang, R.M. Hanson, M. Tobias, S. Krishnan, W.W. Li, J.E. Nielsen, N.

A. Baker, Web servers and services for electrostatics calculations with APBS and

PDB2PQR, J. Comput. Chem. 32 (2011) 1488 –1491 .

¨

F. Johansson, E.J. Homan, O. Loseva, L. Brautigam, L. Johansson, A. Hoglund,

A. Hagenkort, T. Pham, M. Altun, F.Z. Gaugaz, S. Vikingsson, B. Evers,

¨

M. Henriksson, K.S.A. Vallin, O.A. Wallner, L.G.J. Hammarstrom, E. Wiita,

[45] T.J. Dolinsky, P. Czodrowski, H. Li, J.E. Nielsen, J.H. Jensen, G. Klebe, N.A. Baker,

PDB2PQR: expanding and upgrading automated preparation of biomolecular

structures for molecular simulations, Nucleic Acids Res. 35 (Web Server) (2007)

I. Almlof, C. Kalderen, H. Axelsson, T. Djureinovic, J.C. Puigvert, M. Haggblad,

F. Jeppsson, U. Martens, C. Lundin, B.o. Lundgren, I. Granelli, A.J. Jensen,

P. Artursson, J.A. Nilsson, P. Stenmark, M. Scobie, U.W. Berglund, T. Helleday,

MTH1 inhibition eradicates cancer by preventing sanitation of the dNTP pool,

Nature 508 (7495) (2014) 215 –221 .

[46] T.A. Binkowski, S. Naghibzadeh, J. Liang, CASTp: computed atlas of surface

topography of proteins, Nucleic Acids Res. 31 (2003) 3352 –3355 .

20

C. Peng et al.

Bioorganic Chemistry 110 (2021) 104813

[47] D.R. Rhodes, J. Yu, K. Shanker, N. Deshpande, R. Varambally, D. Ghosh,

T. Barrette, A. Pandey, A.M. Chinnaiyan, ONCOMINE: a cancer microarray

database and integrated data-mining platform, Neoplasia 6 (2004) 1 –6 .

[48] A.L. Hopkins, C.R. Groom, A. Alex, Ligand efﬁciency: a useful metric for lead

selection, Drug Discov. Today 9 (10) (2004) 430 –431 .

[60] T. Tsuzuki, A. Egashira, H. Igarashi, T. Iwakuma, Y. Nakatsura, Y. Tominaga,

H. Kawate, K. Nakao, K. Nakamura, F. Ide, S. Kura, Y. Nakabeppu, M. Katsuki,

T. Ishikawa, M. Sekiguchi, Spontaneous tumorigenesis in mice defective in the

MTH1 gene encoding 8-oxo-dGTPase, Proc. Natl. Acad. Sci. U. S. A. 98 (2001)

11456 –11461 .

[49] S. Geschwindner, J. Ulander, P. Johansson, Ligand Binding Thermodynamics in

Drug Discovery: Still a Hot Tip? J. Med. Chem. 58 (2015) 6321 –6335 .

[50] X. Liu, D.C. Speckhard, T.R. Shepherd, Y.J. Sun, S.R. Hengel, L. Yu, C.A. Fowler,

L. Gakhar, E.J. Fuentes, Distinct Roles for Conformational Dynamics in Protein-

Ligand Interactions, Structure 24 (2016) 2053 –2066 .

[61] M. Xu, Q. Zhao, D. Shao, H. Liu, J. Qi, C. Qin, Chenodeoxycholic Acid Derivative

HS-1200 Inhibits Hepatocarcinogenesis and Improves Liver Function in

Diethylnitrosamine-Exposed Rats by Downregulating MTH1, Biomed Res. Int. 2017

(2017) 9 .

[62] T.M. McPhillips, S.E. McPhillips, H.-J. Chiu, A.E. Cohen, A.M. Deacon, P.J. Ellis,

E. Garman, A. Gonzalez, N.K. Sauter, R.P. Phizackerley, Blu-Ice and the Distributed

Control System: software for data acquisition and instrument control at

macromolecular crystallography beamlines, J. Synchrotron Radiat. 9 (2002)

401 –406 .

[51] M. Schauperl, P. Czodrowski, J.E. Fuchs, R.G. Huber, B.J. Waldner, M. Podewitz,

C. Kramer, K.R. Liedl, Binding Pose Flip Explained via Enthalpic and Entropic

Contributions, J. Chem. Inf. Model. 57 (2017) 345 –354 .

[52] A.I. Dragan, C.M. Read, C. Crane-Robinson, Enthalpy-entropy compensation: the

role of solvation, Eur. Biophys. J. 46 (4) (2017) 301 –308 .

[63] Z. Otwinowski, W. Minor, Processing of X-ray diffraction data collected in

oscillation mode, Methods Enzymol. 276 (1997) 307 –326 .

[53] T.T. Talele, The “Cyclopropyl Fragment ” is a Versatile Player that Frequently

Appears in Preclinical/Clinical Drug Molecules, J. Med. Chem. 59 (2016)

8712 –8756 .

[64] A.J. McCoy, R.W. Grosse-Kunstleve, P.D. Adams, M.D. Winn, L.C. Storoni, R.

J. Read, Phaser crystallographic software, J. Appl. Crystallogr. 40 (2007) 658 –674.

`

[54] F. Lovering, J. Bikker, C. Humblet, Escape from ﬂatland: increasing saturation as

an approach to improving clinical success, J. Med. Chem. 52 (2009) 6752 –6756 .

[55] T. Fujishita, T. Okamoto, T. Akamine, S. Takamori, K. Takada, M. Katsura,

G. Toyokawa, F. Shoji, M. Shimokawa, Y. Oda, Y. Nakabeppu, Y. Maehara,

Association of MTH1 expression with the tumor malignant potential and poor

prognosis in patients with resected lung cancer, Lung Cancer 109 (2017) 52 –57.

[56] C. Jüngst, B. Cheng, R. Gehrke, V. Schmitz, H.D. Nischalke, J. Ramakers,

P. Schramel, P. Schirmacher, T. Sauerbruch, W.H. Caselmann, Oxidative damage is

increased in human liver tissue adjacent to hepatocellular carcinoma, Hepatology

39 (6) (2004) 1663 –1672 .

[65] A.W. SchuEttelkopf, D.M. Van Aalten, PRODRG: a tool for high-throughput

crystallography of protein –ligand complexes, Acta Crystallogr. D 60 (2004)

1355 –1363 .

[66] P.D. Adams, P.V. Afonine, G. Bunkoczi, V.B. Chen, I.W. Davis, N. Echols, J.

J. Headd, L.W. Hung, G.J. Kapral, R.W. Grosse-Kunstleve, A.J. McCoy, N.

W. Moriarty, R. Oeffner, R.J. Read, D.C. Richardson, J.S. Richardson, T.

C. Terwilliger, P.H. Zwart, PHENIX: a comprehensive Python-based system for

macromolecular structure solution, Acta Crystallogr. D 66 (2010) 213 –221 .

[67] P. Emsley, K. Cowtan, Coot: model-building tools for molecular graphics, Acta

Crystallogr. D 60 (2004) 2126 –2132 .

[57] Z. Liu, L.-E. Wang, S.S. Strom, M.R. Spitz, R.J. Babaian, J. DiGiovanni, Q. Wei,

Overexpression of hMTH in peripheral lymphocytes and risk of prostate cancer: a

case-control analysis, Mol. Carcinog. 36 (3) (2003) 123 –129 .

[68] P.V. Afonine, R.W. Grosse-Kunstleve, N. Echols, J.J. Headd, N.W. Moriarty,

M. Mustyakimov, T.C. Terwilliger, A. Urzhumtsev, P.H. Zwart, P.D. Adams,

Towards automated crystallographic structure reﬁnement with phenix.reﬁne, Acta

Crystallogr. D 68 (2012) 352 –367 .

[58] C.H. Kennedy, H.I. Pass, J.B. Mitchell, Expression of human MutT homologue

(hMTH1) protein in primary non-small-cell lung carcinomas and histologically

normal surrounding tissue, Free Radic. Biol. Med. 34 (2003) 1447 –1457 .

[59] P. Liu, L.W. Zhao, J. Pol, S. Levesque, A. Petrazzuolo, C. Pﬁrschke, C. Engblom,

S. Rickelt, T. Yamazaki, K. Iribarren, L. Senovilla, L. Bezu, E. Vacchelli, V. Sica,

A. Melis, T. Martin, X. Lin, H. Yang, Q.Q. Li, J.F. Chen, S. Durand, F. Aprahamian,

D. Lefevre, S. Broutin, A. Paci, A. Bongers, V. Minard-Colin, E. Tartour, L. Zitvogel,

L. Apetoh, Y.T. Ma, M.J. Pittet, O. Kepp, G. Kroemer, Crizotinib-induced

immunogenic cell death in non-small cell lung cancer, Nat. Commun. 10 (2019)

1486 .

[69] W. DeLano. The PyMOL Molecular Graphics System, DeLano Scientiﬁc LLC, Palo

Alto, CA, 2008 .

[70] C.H. Ho, S.P. Liu, L.C. Hsu, W.L. Chang, C.C.K. Chao, J.H. Guh, Repurposing of

Phentolamine as a Potential Anticancer Agent Against Human Castration-Resistant

Prostate Cancer: A Central Role on Microtubule Stabilization and Mitochondrial

Apoptosis Pathway, Prostate 75 (2015) 1454 –1466 .

[71] M.C. Lin, J.J. Lin, C.L. Hsu, H.F. Juan, P.J. Lou, M.C. Huang, GATA3 interacts with

and stabilizes HIF-1alpha to enhance cancer cell invasiveness, Oncogene 36 (2017)

4243 –4252 .

21

Article Doi

Inhibitor development of MTH1 via high-throughput screening with fragment based library and MTH1 substrate binding cavity

DOI: 10.1016/j.bioorg.2021.104813

Source and publish data:

Authors:

Article abstract of DOI:10.1016/j.bioorg.2021.104813

Full text of DOI:10.1016/j.bioorg.2021.104813

Products guided by the article

R&D Labs maybe for 1005-37-4

Relevant to this article

Hot Product