Chemical Inhibition of Human Thymidylate Kinase and Structural Insights into the Phosphate Binding Loop and Ligand-Induced Degradation

Article abstract of DOI:10.1021/acs.jmedchem.6b01280

Targeting thymidylate kinase (TMPK) that catalyzes the phosphotransfer reaction for formation of dTDP from dTMP is a new strategy for anticancer treatment. This study is to understand the inhibitory mechanism of a previously identified human TMPK (hTMPK) inhibitor YMU1 (1a) by molecular docking, isothermal titration calorimetry, and photoaffinity labeling. The molecular dynamics simulation suggests that 1a prefers binding at the catalytic site of hTMPK, whereas the hTMPK inhibitors that bear pyridino[d]isothiazolone or benzo[d]isothiazolone core structure in lieu of the dimethylpyridine-fused isothiazolone moiety in 1a can have access to both the ATP-binding and catalytic sites. The binding sites of hTMPK inhibitors were validated by photoaffinity labeling and mass spectrometric studies. Taking together, 1a and its analogues stabilize the conformation of ligand-induced degradation (LID) region of hTMPK and block the catalytic site or ATP-binding site, thus attenuating the ATP binding-induced closed conformation that is required for phosphorylation of dTMP.

Full text of DOI:10.1021/acs.jmedchem.6b01280

Subscriber access provided by CORNELL UNIVERSITY LIBRARY
Article
Chemical Inhibition of Human Thymidylate Kinase and Structural Insights
into the Phosphate Binding Loop and Ligand-induced Degradation
Yi-Hsuan Chen, Hua-Yi Hsu, Ming-Tyng Yeh, Chen-Cheng Chen, Chang-Yu Huang, Ying-Hsuan
Chung, Zee-Fen Chang, Wei-Chen Kuo, Nei-li Chan, Jui-Hsia Weng, Bon-chu Chung, Yu-Ju
Chen, Cheng-Bang Jian, Ching-Chieh Shen, Hwan-Ching Tai, Sheh-Yi Sheu, and Jim-Min Fang
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b01280 • Publication Date (Web): 17 Oct 2016
Downloaded from http://pubs.acs.org on October 20, 2016
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society.
1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the
course of their duties.

Page 1 of 61
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Chemical Inhibition of Human Thymidylate Kinase and Structural Insights into the
Phosphate Binding Loop and Ligand-induced Degradation
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1,§
2,§
1
3
3
Yi-Hsuan Chen, Hua-Yi Hsu, Ming-Tyng Yeh, Chen-Cheng Chen, Chang-Yu Huang ,
3
3
4
4
5
1
Ying-Hsuan Chung , Zee-Fen Chang, Wei-Chen Kuo, Nei-Li Chan, Jui-Hsia Weng,
5
1,6
1
1
Bon-chu Chung, Yu-Ju Chen, Cheng-Bang Jian, Ching-Chieh Shen, Hwan-Ching Tai,
2
,7,
1,8,
Sheh-Yi Sheu, * and Jim-Min Fang. *
1
Department of Chemistry, National Taiwan University, Taipei 106, Taiwan.
2
Department of Life Sciences and Institute of Genome Sciences, National Yang-Ming
University, Taipei 112, Taiwan.
3
4
Graduate Institute of Biochemistry and Molecular Biology, National Yang-Ming University,
Taipei 112, Taiwan.
Institute of Biochemistry and Molecular Biology, College of Medicine, National Taiwan
University, Taipei 100, Taiwan.
5
6
7
8
Institute of Molecular Biology, Academia Sinica, Taipei 115, Taiwan.
Institute of Chemistry, Academia Sinica, Taipei, 115, Taiwan.
Institute of Biomedical informatics, National Yang-Ming University, Taipei 112, Taiwan.
The Genomics Research Center, Academia Sinica, Taipei 115, Taiwan.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 61
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
§
Y.-H.C. and H.-Y.H. contributed equally to this work.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ACS Paragon Plus Environment

Page 3 of 61
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
ABSTRACT
Targeting thymidylate kinase (TMPK) that catalyzes the phospho-transfer reaction for
formation of dTDP from dTMP is a new strategy for anti-cancer treatment. This study is to
understand the inhibitory mechanism of a previously identified human TMPK (hTMPK)
inhibitor YMU1 (1a) by molecular docking, isothermal titration calorimetry, and photoaffinity
labeling. The molecular dynamics simulation suggests that 1a prefers binding at the catalytic
site of hTMPK, whereas the hTMPK inhibitors that bear pyridino[d]isothiazolone or
benzo[d]isothiazolone core structure in lieu of the dimethylpyridine-fused isothiazolone
moiety in 1a can have access to both the ATP-binding and catalytic sites. The binding sites of
hTMPK inhibitors were validated by photoaffinity labeling and mass spectrometric studies.
Taking together, 1a and its analogues stabilize the conformation of ligand-induced
degradation (LID) region of hTMPK and block the catalytic site or ATP-binding site, thus
attenuating the ATP binding-induced closed conformation that is required for phosphorylation
of dTMP.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Keywords:
Thymidylate kinase; Inhibitors; Molecular modeling; Photoaffinity probe; Isothermal titration
calorimetry; Mass spectrometry.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 61
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
INTRODUCTION
The enzymes of thymidylate synthesis have been targets for the actions of many
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
anti-cancer drugs for more than 60 years. However, human thymidylate kinase (TMPK, EC
2
.7.4.9) has been a missing target, until our recent finding that blocking TMPK is able to
2
sensitize cancer cells for low-dose doxorubicin chemotherapy. Furthermore, the aggressive
LKB-deficient lung cancers are hypersensitive to TMPK inhibition, suggesting the therapeutic
3
applicability of TMPK inhibitors. TMPK is a homodimeric enzyme that catalyzes the
2
+
4
phosphorylation of dTMP to dTDP in the presence of Mg ion and ATP. According to the
5
–7
sequence and structure analyses, the phosphate binding loop (P-loop), DR(Y/H) motif and
ligand-induced degradation (LID) loop are highly conserved and essential for the function of
8
TMPK. P-loop binds the α- and β-phosphoryl groups of ATP. The Asp96 residue of DR(Y/H)
2
+
motif binds the ATP–Mg complex, while the Arg97 residue is involved in bringing ATP and
dTMP together to induce a phosphoryl group transfer. The LID region is a flexible stretch
nearby ATP binding site; however, its function in the connection of ATP (the phosphate donor)
8
and dTMP (the phosphate acceptor) is unclear.
The importance of TMPK in living organisms is attributed to its situated position at the
junction of the de novo and salvage pathways of dTTP biosynthesis. TMPK is considered as a
9
potential therapeutic target for various disease treatments. For example, dTMP analogs have
been synthesized as the inhibitors against the TMPK of Mycobacterium tuberculosis
ACS Paragon Plus Environment

Page 5 of 61
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
0
(
TMPKmt) to suppress M. tuberculosis infection. Interestingly, these inhibitors have no
10
inhibitory effect on human TMPK (hTMPK). It is noted that the sequence variation
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
surrounding the three essential loops confers inhibitor selectivity, even though the common
5, 11, 12
folding structures of TMPK from different species are similar.
We have previously shown that blocking hTMPK is able to sensitize cancer cells for
low-dose doxorubicin chemotherapy, and identified YMU1 (1a) as a specific hTMPK
2
inhibitor (Chart 1). In contrast, 1a has no inhibitory effect on the activity of purified
thymidine kinase 1 (TK1) that is the enzyme responsible for the conversion of thymidine to
2
dTMP. The mechanistic investigation demonstrates that the lack of hTMPK functionality in
cancer cells leads to dUTP misincorporation in DNA repair, resulting in cancer cell death.
2
Overexpression of TMPK can rescue the DNA lesions caused by 1a. The conjunction
treatment of 1a with low dose of doxorubicin suppresses tumor growth in vitro and in vivo.
Compound 1a represents a new class of TMPK inhibitor since it is not a dTMP or nucleoside
13
analog. The kinetics study of TMPK revealed that 1a is a mixed competitive inhibitor;
2
however the detailed binding mode of 1a has not yet resolved. Understanding how 1a
inhibits hTMPK would provide useful information critical for optimization of hTMPK
inhibitors.
Crystallographic analysis of the hTMPK–1a complex, either by co-crystallization or by
soaking the pre-grown hTMPK crystals with 1a, proved to be difficult. Thus, molecular
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 61
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
docking and probe labeling are two alternative approaches to study the binding sites of
hTMPK. In this study, an ATP-free structure of hTMPK was used to perform molecular
modeling experiments for understanding the molecular mechanism underlying hTMPK
inhibition by 1a and its arene-fused isothiazolone analogs 1b–3b (Chart 1). A photoaffinity
probe 4 bearing moieties of arylazide and biotin (Chart 1) was designed to identify the
potential inhibitor binding sites of hTMPK.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Chart 1. Chemical structures of 1a, arene-fused isothiazolone analogues 1b–3b, and
photoaffinity probe 4.
RESULTS AND DISCUSSION
The crystal structure of ATP-free form of hTMPK is constructed with the LID
region for molecular simulations. Understanding the interaction of hTMPK with its
inhibitors is critical for the development of pharmaceutical agents. Toward this aim, molecular
simulations were performed to examine these interactions, and the important features for
hTMPK–inhibitor recognition were described. We used the molecular mechanics/Poisson–
ACS Paragon Plus Environment

Page 7 of 61
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
4
Boltzmann Surface Area method (MM/PBSA) to estimate the enzyme–inhibitor binding
affinity by looking at the free energy change of the system. We were unable to obtain
diffracting crystals of hTMPK in complex with ligand, reflecting the fact that large structural
rearrangements would be required to accommodate inhibitor access to the enzyme active site.
Our kinetic analysis has shown that preincubation of 1a reduces the ATP binding affinity of
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
hTMPK; therefore, the structure of an ATP-free form of hTMPK appears to be a more
suitable starting point for structural modeling of hTMPK–inhibitor complexes. However, the
LID region was not resolved in the crystal structure of ATP-free hTMPK, therefore we
8
constructed the LID region based on the loop position in ADP-bound hTMPK (PDB 1E2D)
(Figure 1A). A comparison of the structures between the ATP-free and ADP-bound hTMPK
was shown in Figure 1A; in ATP-free hTMPK, the modeled LID region was followed by 20
ns molecular dynamics (MD) simulation to obtain a relaxed conformation of the LID region
(
Figure 1C).
The analyses of the hydrogen-bonding network with the residues between the P-loop and
the LID region are shown for the ADP-bound (Figure 1B) and the ATP-free hTMPKs
followed by MD simulation (Figure 1D). This analysis points to the existence of a
hydrogen-bonding network within the ATP binding site, consisting of residues Asp15, Arg16,
Ala17, Arg143, Gly144, Glu152, and Ala180 (Figure 1B); the ATP binds in a cleft lined by
Arg16, Gly18, Lys19, Ser20, Thr21, Arg143, and Lys182. A distinct conformation between
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 61
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ADP-bound and ATP-free forms is the P-loop. In ADP-bound form, Arg16 points to the LID
region and forms hydrogen-bonds with Gly144, Ala145, and Glu152; the carboxyl group of
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2+
Asp15 and the phosphate groups of ADP and TMP coordinates with Mg (Figure 1B). In
ATP-free form, the residues of Asp15, Arg16, Ala17, and Glu149 form hydrogen-bonding
network; the side chain of Arg16 points to the phosphate group of TMP and the carboxyl
2
+
groups of Asp15 and Glu152 chelate Mg ion (Figure 1D). Due to the loss of the ATP
interaction, the LID region becomes more flexible after MD simulation.
Figure 1. Structural model of hTMPK. (A) Overlap of the structures of ADP-bound hTMPK
(
yellow) and the ATP-free hTMPK (gray). Residues of the P-loop (13–17) are shown in ball
ACS Paragon Plus Environment

Page 9 of 61
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
and stick representation and colored in blue; the LID region (135–150) can be seen in
magenta; the DR motif (Asp96/Arg97) is shown in red. (B) Close-up view of the ADP binding
site with residues shown. TMP is shown in stick representation with its color based on atom
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2+
type, carbon (green), oxygen (red), nitrogen (blue) and hydrogen (white); Mg ion is shown
in pink. The hydrogen-bonds are shown in dash lines. (C) Depiction of the structures of the
relaxed ATP-free TMPK (gray) and the ADP-bound hTMPK (yellow). Four possible ligand
docking sites are denoted as A, B, C and D sites, which were created using AutoDock Vina
1
5
program . (D) Close-up view of ATP-free hTMPK structure with relaxed LID region and
residues shown.
Molecular docking analysis with 1a reveals four binding sites in hTMPK. We then
performed the docking analysis with compound 1a to screen possible binding sites. The
results revealed four related binding sites for 1a in Figure 1C, where A-site is above the LID
region, B-site is surrounding the catalytic pocket between P-loop and LID region, C-site is
beneath and surrounding TMP, and D-site is ATP binding site. When both ATP and dTMP
exist, an ATP molecule will reside at the D-site and a dTMP molecule will reside at the C-site.
During the phospho-transfer process, the ATP molecule will be relocated from D-site to the
8
catalytic pocket (B-site) to furnish the phosphorylation reaction of dTMP. The docking scores
show that A- and C-sites are not good for ligand binding, while B- and D-sites have higher
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 61
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
docking scores. Therefore, both B- and D-sites are used to investigate the inhibitory
mechanism at the molecular level.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
In the catalytic pocket (B-site), the optimized structure of the hTMPK–1a complex
subjected to 35 ns MD simulation is shown (Figure 2A). Structural properties were monitored
by computing the root-mean-square deviation (RMSD) between the snapshots during the MD
simulation and the starting coordinates to ascertain that the simulations were stable and
converged. The RMSD computed for the complex hTMPK–1a backbone atoms was very
stable with the value below 2.5 Å (Figure S1A in Supporting Information (SI)). The residues
Ala141, Ala145, Phe146, Ala140, Leu137, Leu135, Ala17, and Val14 form a hydrophobic
pocket surrounding compound 1a. We found that the flexibility of the LID region is dramatic
in the ligand-free hTMPK. In contrast, in the hTMPK–1a complex, 1a interacts with LID
region and forms hydrogen-bonds with the amide N atoms of Ala145 and Glu149 to slow
down the motion of LID region (Figure S1B and S1C in SI). Therefore, the molecular
mechanism of 1a in inhibition of hTMPK is considered to disrupt the hydrogen-bonding
network between LID region and P-loop and blocks the catalytic site, after which ATP binding
is unable to induce a closed conformation required for the catalytic reaction of hTMPK.
Molecular docking analyses predict compounds 2b and 3b are able to bind at either
the ATP-binding or catalytic sites of hTMPK. According to our hTMPK inhibition assays,
ACS Paragon Plus Environment

Page 11 of 61
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
compounds 1a, 1b and 1c that bear the end groups (CO Et, CO t-Bu and H) in different sizes
2
2
did not have appreciable effect on the inhibitory activity. According to data of 1a with
hTMPK, we thus undertook the molecular docking analyses of compounds 2b and 3b to
compare their binding modes with that of 1a in hTMPK. Instead of the ethoxycarbonyl group
in 1a, both 2b and 3b have a bulkier tert-butoxycarbonyl (Boc) group at the end. Compound
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
b modifies the dimethylpyridine-fused isothiazolone moiety in 1b by deletion of the two
methyl substituents, whereas 3b further changes the core structure to benzo[d]isothiazolone.
The binding modes of 1a, 2b and 3b to hTMPK are shown in Figure 2A–F. The binding free
energies of three ligands and the assay results are listed in Figure 2G, while the detailed
energy components are given in SI (Table S1).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 61
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 2. Binding modes of hTMPK–ligand complexes and the assay results. (A) compound
1a in the catalytic site, (B) compound 2b in the catalytic site, (C) compound 3b in the
catalytic site, (D) compound 1a in the ATP-binding site, (E) compound 2b in the ATP-binding
site, and (F) compound 3b in the ATP-binding site. All notations are the same as the legend of
Figure 1. Hydrogen bond is shown in blue (dash line). The structure was taken from the last
3
5 ns MD trajectory. (G) Ligand binding free energy (ΔGbinding), IC and EC values against
50 50
a
b
hTMPK. Errors are given as standard deviation (SD). Ligand binding free energies are
c
d
calculated according to Eq. [1] (see Experimental section). N.D.: not detected. IC values
5
0
e
are derived from luciferase-coupled TMPK assay in triplicate independent experiments. EC
50
vales are derived from viability assay of 32D/Bcr-Abl cancer cells in triplicate independent
experiments.
ACS Paragon Plus Environment

Page 13 of 61
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
The binding mode of 2b at the catalytic site (Figure 2B) is similar to that of 1a (Figure
2
A), stabilizing Ala145 and Glu149 through hydrogen bonding with the C carbonyl and the
3
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
N atom of the heterocyclic core structure. However, the binding free energy (ΔG
= –
7
binding
2
3.3 kcal/mol) of 2b at the B-site is still weaker than that of 1a (ΔGbinding = –39.8 kcal/mol)
(Figure 2G). When 3b was superimposed into the catalytic site (B-site) based on the position
of 1a, there is no hydrogen bond between hTMPK and 3b, thus the binding free energy is
further decreased (ΔGbinding = –16.2 kcal/mol). These results indicate that the pyridine moiety
of 1a and 2b may be superior to the benzene moiety of 3b to exert hydrogen bonds for better
binding affinity in the B-site of TMPK.
In sharp contrast, the MD simulation shows that 1a is unable to reside at the ATP-binding
site (D-site), presumably because the methyl groups of the heterocyclic core would cause
severe hindrance (Figure 2D). In comparison, 2b can form hydrogen bonds via
N ∙∙∙∙HN(Arg143) and O ∙∙∙∙HO (Thr21) at the D-site (Figure 2E). Compound 3b also prefers
7
3
β
to stay at the D-site during MD simulation through hydrophobic interactions (Figure 3F).
Taken together the MD simulations at B- and D-sites, a ligand (e.g. 2b and 3b) capable of
blocking both sites may become a stronger TMPK inhibitor than 1a that only adapts a single
B-site. This deduction is supported by measurement of the inhibitory activity against hTMPK
(Figure 2G).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 61
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Synthesis of compounds for evaluation of hTMPK inhibitory activity. To verify the
binding affinity predicted by the MD simulation, compounds 1a–3b were synthesized and the
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
inhibitory activities against hTMPK were measured (Figure 2G). The synthesis of compound
2
1
a has been reported, and the analogous compounds 1b, 2b and 3b were synthesized by the
similar procedures (Scheme 1). For example, 2-chloronicotinonitrile was reacted with
20
thiourea to give 2-mercaptonicotinonitrile (5b), which was then treated in concentrated
2
H SO to give pyridine-fused isothiazolone 6b as the oxidative cyclization product. The
2
4
reaction of 6b with a relatively soft electrophile of iodine reagent 7 afforded the N-alkylation
product 2b, predominating over the O-alkylation. Compound 1c was obtained by removing
the Boc group in 1b with trifluoroacetic acid (TFA).
Scheme 1. Synthesis of hTMPK inhibitors 1b–3b.
Reagents and conditions: (i) conc. H SO , reflux, 4–5 h; 6a, 61%; 6b, 30%. (ii) Cs CO , Et N,
2
4
2
3
3
CH Cl , rt, 5–24 h; 1b, 59%; 2b, 40%; 3b, 60%. (iii) TFA, CH Cl , rt, 0.5 h; 100%.
2
2
2
2
ACS Paragon Plus Environment

Page 15 of 61
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Photoaffinity labeling provides a directly way to identify the binding domain within the
16
target protein. To design a photoaffinity probe, structural modification of pharmaceutical
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
17
skeleton is needed by incorporation of photosensitive group and reporter unit. We chose
18
aryl azide as the photoreactive group and a biotin tag to merge with the structural skeleton
of the potent compound 3b. Thus, the photoaffinity probe 4 was constructed to identify the
interactions between hTMPK and this new class of inhibitors. Furthermore, purification of
1
9
the covalently bound hTMPK–4 complex would be facilitated by the biotin tag. Scheme 2
shows the synthesis of probe 4. The substitution reaction of 4-bromo-2-fluorobenzonitrile
with Na S occurred selectively via an addition–elimination mechanism to give
2
2
0
4-bromo-2-mercaptobenzonitrile (5d), which underwent an oxidative cyclization reaction in
2
concentrated H SO to give benzoisothiazolone 6d. The N-alkylation of 6d with iodine
2
4
reagent 7 was carried out to afford 8. The reaction of aryl bromide 8 with NaN in the
3
presence of CuI, sodium ascorbate and N,N′-dimethylethylenediamine (DMEDA) gave the
21
desired azido product 10, albeit in low yield (12%). The major product turned out to be the
aniline 9 (58%), though the mechanism for its formation was unclear. It has been suggested
that azidobenzene may undergo in situ reduction to aniline in the reaction conditions
22
containing CuI and NaN . Fortunately, aniline 9 could be treated with NaNO and NaN in
3
2
3
acidic condition to give the desired azido compound 10, presumably via Sandmeyer reaction
2
3
of the intermediate diazonium salt. The Boc protecting group in 10 was removed by TFA to
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 61
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
give amine 11, which was then linked to a biotin labeling reagent 12 to furnish the
24
photoaffinity probe 4.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 2. Synthesis of probe 4.
Reagents and conditions: (i) conc. H SO , reflux, 5 h; 94%. (ii) i-Pr NEt, THF, rt, 4 h; 41%.
2
4
2
o
(
iii) NaN , sodium ascorbate, CuI, DMEDA, EtOH, H O, 100 C, 2 h; 58%. (iv) NaNO ,
3
2
2
NaN , N-methyl-2-pyrrolidonium bisulfate, H O, 1 h; 45%. (v) CF CO H, CH Cl , 30 min;
3
2
3
2
2
2
8
7%. (vi) Et N, DMSO, 8 h; 40%. DMEDA is N,N′-dimethylethylenediamine.
3
hTMPK inhibition assay supports that compounds 2b and 3b are stronger
2
5
inhibitors than 1a. A luciferase-coupled TMPK assay was applied to evaluate the
inhibitory activity of 1a–4 against hTMPK. Our previous study has indicated that compound
1a is an effective inhibitor against hTMPK, showing 70.0 ± 7.0% inhibition at a
ACS Paragon Plus Environment

Page 17 of 61
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
concentration of 2 μM. Compounds 1b and 1c at 2 μM also showed similar inhibitory
activities (76.3 ± 2.5% and 73.4 ± 5.7%) regardless the change of end R groups (CO Et in 1a,
2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
CO t-Bu in 1b and H in 1c). In comparison, change of the heterocyclic core structure caused
2
remarkable effect as shown by much stronger hTMPK inhibition of compounds 2b and 3b
(91.3 ± 4.0% and 91.3 ± 5.7% at 2 μM). Probe 4 at 2 μM inhibited 86.9 ± 0.7% of hTMPK
activity, indicating that the small azido substituent and the long-chain end group did not
interfere with the binding in hTMPK. The IC values of 1a, 2b and 3b were determined to be
50
1.65 ± 1.1, 0.47 ± 0.08 and 0.35 ± 0.01 μM by luciferase-coupled TMPK assay, respectively
(Figure 2G), indicating that 2b and 3b are stronger binders than 1a. The conventional
NADH-coupled assay also confirmed that compounds 2b and 3b had higher hTMPK
inhibitory activity than 1a (Figure S2 in SI). As suggested by the MD simulations (Figure 2),
a compound can bind at either the B- or D-site to suppress the activity of hTMPK. In
comparison, 1a can only have access to B-site, while compounds 2b and 3b can be
incorporated into both the B- and D-sites of hTMPK. The higher inhibitory activity of 2a and
3
b may be related to the higher probability of access to hTMPK.
The ability of these compounds on suppressing viability in normal versus cancer cells
was further compared. To this end, we chose 32D myeloid progenitor cells as normal and
Bcr-Abl transformed 32D as cancer cells. Bcr-Abl is a hallmark of chronic myeloid leukemia
due to chromosomal reciprocal translocation between chromosome 22 and chromosome 9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 61
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(
t(9;22)(q34;q11)). The resulting Bcr-Abl fusion protein transforms normal hematopoietic
26
cells into leukemic cells . We treated 32D and 32D/Bcr-Abl- cells with 1a, 2b and 3b for
viability determination. Treatment with compounds 1a, 2b and 3b up to 2 μM for 3 days did
not cause appreciable effect on 32D cells (Figure S3 in SI). In contrast, 1a, 2b and 3b were
toxic to 32D/Bcr-Abl cells, showing EC50 values of 0.88, 1.40 and 0.65 M, respectively
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(Figure 2G). Compared to 1a and 2b, the benzo[d]isothiazolone 3b was more effective and
selective in inhibiting cell proliferating in Bcr-Abl transformed leukemia cells.
Isothermal titration calorimetry reveals that the hTMPK–inhibitor complex is
formed in 1:1 stoichiometry. Isothermal titration calorimetry (ITC) was performed to
elucidate the binding mode of inhibitor with hTMPK. In ITC experiment, it is necessary to
have the test compound and protein solubilized in the same solution. Both compound 1c
(containing an amine moiety) and hTMPK are soluble in PBS buffer (pH = 7.4), and thus
suitable for the ITC experiment. In contrast, compounds 2b and 3b cannot dissolve well in
PBS buffer without addition of DMSO. As compound 1c still has an appreciable inhibitory
activity against hTMPK, we thus chose 1c to perform the ITC analysis (Figure 3). The ITC
experiment confirms that 1c binds with hTMPK to form an hTMPK–1c complex in 1:1
–1
stoichiometry with the binding constant of 127,000 ± 19,600 M in PBS solution. The
disposition of compound 1c in the active site of hTMPK is thermodynamically favored as
ACS Paragon Plus Environment

Page 19 of 61
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
shown by the negative free-energy change that is mainly attributed to the enthalpy change
(ΔH = –6.95 kcal/mol) with very little entropy change (ΔS = 0.047 cal/mol/deg).
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 3. Isothermal titration calorimetry (ITC) experiment. ITC shows that hTMPK and
compound 1c forms a complex of 1:1 stoichiometry in PBS solution.
Amino acid residues in proximity of the ATP-binding site of hTMPK are identified
by photoaffinity labeling and mass spectrometric analysis. Probe 4 containing a
photoreactive arylazide moiety and a biotin affinity handle was preincubated with purified
o
hTMPK protein in PBS buffer at 25 C for 30 min, and then subjected to irradiation with
27
2
54-nm UV light (Figure 4A). The highly reactive nitrene intermediate was expected to
cross link with hTMPK by forming covalent bonds with the proximate amino acid residues in
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 61
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
18
the active site. The covalent attachment of probe 4 to hTMPK was visualized by
streptavidin–horseradish peroxidase (HRP) in Western blot analysis (Figure 4B). The
photolabeling experiment was also carried out without preincubation. In the absence of UV
light, we did not observe non-specific labeling of hTMPK (Figure 4B).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ACS Paragon Plus Environment

Page 21 of 61
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. Photoaffinity probe labeling. (A) Schematic for photoaffinity labeling. (B) Western
blot of hTMPK photolabeled with probe 4, visualized using streptavidin–horseradish
peroxidase. (C) Sequences of photolabeled peptides. Photolabeled hTMPK was separately
digested with trypsin and chymotrypsin, and biotinylated peptides were enriched by affinity
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 61
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
chromatography and analyzed by LC–MS/MS. Tryptic and chymotryptic peptides showing
probe conjugation are marked by red and orange boxes, respectively. (D) & (E) Positions of
photolabeled peptides mapped to hTMPK structure (PDB: 1E2D, no coordinates available for
N-terminal residues MAA), with tryptic peptides marked in red (D) and chymotryptic peptide
in orange (E). The D-site is marked, and the black and grey arrows indicate two general
binding regions of the probe 4.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The modification sites of hTMPK by probe 4 were investigated by a mass spectrometric
approach. We utilized two different proteases, trypsin and chymotrypsin, to separately digest
photolabeled hTMPK. By affinity chromatography with immobilized monomeric avidin, the
photolabeled peptides bearing terminal biotin moiety were enriched and then subjected to
LC–MS/MS analysis (Figure 4C). The ester linkage in probe 4 was sometimes hydrolyzed
under the acidic condition of reversed-phase chromatography, and hence the mass shift of
labeled peptides was either +648.24 or +422.16 Da (Table S2 in SI). We used both mass shift
and strong enrichment during avidin affinity chromatography as the criteria for identification
of modified peptides. With trypsin and chymotrypsin digestion, we identified three and two
probe-modified peptides, respectively (Figure 4C and Table S2). By mapping these three
modified peptides onto the three dimensional structure of hTMPK, two tryptic (residues 20–
25 and 177–182, Table S2) and one chymotryptic peptide (165–168) appeared at or in the
ACS Paragon Plus Environment

Page 23 of 61
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
vicinity of D-site (indicated by black arrow in Figures 4D and 4E). The extreme N-terminus
and the extreme C-terminus, which were unstructured in crystallography and in close
proximity, were also labeled by probe 4. This probably represented a nonspecific binding site
on the opposite face of the catalytic pocket (indicated by grey arrow in Figures 4D and 4E).
Combining the results of trypsin and chymotrypsin digestion, the primary binding sites of
probe 4 appeared to be the ATP binding site (D-site), consistent with the molecular docking
results of compound 3b (Figure 2), from which probe 4 was derived. Though no
probe-modified peptides for binding of probe 4 at the B-site were found in our experiment,
one should not overlook the possibility that our current experiment might not be sensitive
enough to detect such probe-modified peptides that could occur in small amount or difficult
to ionization in MS analysis.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
TMPK inhibitors 1a–4 are not PAINS. There is a concern about whether compounds
28,29
1
a–4 are pan assay interference compounds (PAINS).
Multiple lines of experimental
2
evidence exclude this possibility. (1) Our previous work has revealed that 1a has no
inhibitory effect on the activity of purified thymidine kinase 1 (TK1) that is the enzyme
responsible for the conversion of thymidine to dTMP. (2) We also demonstrates that the effect
of 1a on impairing DNA repair can be suppressed by three different molecular events that can
2
remove dUTP misincorporation in DNA. These molecular events include overexpression of
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 61
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
dUTPase, knockdown of uracil-DNA glycosylase and disruption of ribonucleotide reductase
recruitment to DNA damage sites. If 1a were a PAIN, the effect of 1a would not have been
overcome by these events. (3) In this study, we further show that 1a and its analogs (2b, 3b
and 4) have no inhibitory effect on tyrosine kinase activity (IC > 100 μM) (Table S3 in SI).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
0
(4) The TLC monitored isotope labeling experiment shows that 1a inhibits the formation of
3
2
P-dTDP by human TMPK but not candida Cdc8, which is a human TMPK ortholog sharing
4
1% identity in amino-acid sequence (Figure S4 in SI). (5) Normal cells (e.g. 32D myeloid
progenitor cells) are not affected by treatment of 1a at 2 μM for 3 days (Figure S3 in SI),
suggesting no appreciable toxicity of 1a. If 1a had nonspecific inhibitory effect, the growth of
normal cells would have been affected. (6) Promiscuous PAINS often contain the functional
groups (e.g., alkyl halide and α,β-unsaturated ester) that are reactive to the nucleophilic
sulfanyl group of Cys residue. However, no protein band corresponding to nonspecific
addition of Cys is observed in Western blot after incubation of probe 4 with hTMPK under
UV-free condition (Figure 4B). (7) The three Cys residues (Cys31, Cys117 and Cys163) in
hTMPK are mapped (Figure S5 in SI), but none of them exist in the ATP-binding site or
catalytic site. Although Cys163 is in the proximity of hTMPK active site, irradiation of probe
4
labels its neighboring region (165–168) but not this particular residue (Figure 4C). (8)
-Phenylbenzo[d]isothiazol-3(2H)-one (structure A in Chart S1 of SI) is considered a PAIN
2
2
9
acting as a covalent modifier of proteins. We thus examined whether compounds 1a and A
ACS Paragon Plus Environment

Page 25 of 61
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
could be attacked by the Cys residues of hTMPK? We chose Ellman′s reagent,
5
,5′-dithiobis(2-nitrobenzoic acid) (DTNB), as a positive control and DMSO vehicle as a
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
o
negative control. After incubation of hTMPK with 1a, A and DTNB at 25 C for 30 min,
respectively, the hTMPK solutions were readily digested by trypsin, without prior reduction
with dithiothreitol (DTT). The LC–MS/MS analyses showed 95–100% coverage of the
hTMPK sequence, and only compounds A and DTNB modified hTMPK by formation of
disulfide bonds with Cys residues (Table S4 in SI). This result confirms that the binding of 1a
with hTMPK is noncovalent and not through disulfide formation. Taking together the
evidence from different enzymatic and cell-based assays, compounds 1a–4 are specific
hTMPK inhibitors, and unlikely belong to the category of PAINS.
We also found that benzoisothiazolone compounds bearing different substituents may
alter their binding modes. For comparison, we prepared compounds 13–17 (Chart S1 in SI)
having the structures related to the hTMPK inhibitors 1a–3b. Their inhibitory activities were
evaluated by luciferase-coupled TMPK assay. Though compound 3a is an effective hTMPK
inhibitor (70% inhibition at 2 μM), compound 13 that is structurally similar to 3a but having
a longer linker showed no appreciable hTMPK inhibitory activity (7% inhibition at 2 μM).
Unlike PAIN compound A, benzoisothiazolone 14 having a 3-chloro-4-fluorophenyl
substituent, instead of the phenyl group in A, did not bind with hTMPK (< 2% inhibition at 2
μM). Unlike 1a, 2b and 3b that were derived from N-alkylation, the corresponding
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 61
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
O-alkylation compounds 15–17 showed low hTMPK inhibition. It is clear that not all
benzoisothiazole or pyridinoisothiazole compounds are PAINS.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
CONCLUSIONS
TMPK is situated at the pivotal position of the de novo and salvage pathways of dTTP
biosynthesis. When the functionality of TMPK is suppressed, the supply of dTTP is reduced,
and dUTP misincorporation can occur in DNA repair to cause cancer cell death. Compound
1a is a specific hTMPK inhibitor to sensitize cancer cells for low-dose doxorubicin
2
chemotherapy. Compound 1a contains the main structural scaffold of pyridine-fused
isothiazolone that belongs to a new class of TMPK inhibitor. To understand the binding mode
of 1a in hTMPK will help to develop this new class of hTMPK inhibitors in cancer treatment.
In this study, we first carry out molecular simulations using the crystal structure of an
ATP-free form of hTMPK with implantation of the LID region to show the important features
and interactions of hTMPK with the arene-fused isothiazolone inhibitors. According to the
results of our molecular docking experiments, hTMPK has an ATP binding site (D-site) and a
B-site surrounding the catalytic pocket between P-loop and LID region to incorporate the
arene-fused isothiazolone inhibitors. The ITC analysis confirms that the arene-fused
isothiazolone inhibitor forms a 1:1 complex with hTMPK. Compound 1a prefers to bind at
the B-site of hTMPK, whereas compounds 2b and 3b can bind at ether the B- or D-sites.
ACS Paragon Plus Environment

Page 27 of 61
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Thus, compounds 2b (IC = 0.47 μM) and 3b (IC = 0.35 μM) exhibit about 4-fold higher
5
0
50
binding affinity than 1a (IC = 1.65 μM). In the cell viability assay, compound 3b inhibits
5
0
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
the proliferation of Bcr-Abl transformed 32D cell with more efficiency but less cytotoxicity
to 32D myeloid progenitor cells compared to compound 1a.
As the end groups in these compounds do not exhibit appreciable effect on the hTMPK
inhibitory activity, the photoaffinity probe 4 is designed to bear a photoreactive moiety of
arylazide and a biotin tag. The photoaffinity labeling experiment works well to tag the
interacting amino acid residues in the active sites of hTMPK. The probe-modified peptides
are identified by mass spectrometric method to show that most of the interacting peptides
occur at the D-site as predicted by our molecular docking experiments. The preferable
location of compounds 2b, 3b and 4 in the ATP-binding site may be attributable to their main
structures of pyridine- and benzene-fused isothiazolones that act as proper mimetics of the
purine moiety of ATP. In comparison, 1a has less inhibitory activity presumably because it
has two methyl groups on the pyridine ring to make severe hindrance against its disposition
in the ATP-binding site. Taking together, our present study provides clear evidence for the
binding mode of the arene-fused isothiazolone inhibitors, which disrupt the connection
between LID region and P-loop that are crucial to the function of hTMPK in phospho-transfer
from ATP to dTMP. As TMPK inhibition can selectively sensitize cancer cells to low-dose
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 61
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
chemotherapy, the structural information of TMPK–inhibitor will be useful for design of
potent inhibitors in various cancer treatments.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
EXPERIMENTAL SECTION
General. All the reagents and solvents were commercially available and used without
further purification unless indicated otherwise. All solvents were anhydrous grade unless
indicated otherwise. Dichloromethane (CH Cl ) was distilled from CaH . All air or moisture
2
2
2
sensitive experiments were performed under nitrogen. Reactions were magnetically stirred
and monitored by thin-layer chromatography (TLC) on 0.25 mm E. Merck silica gel 60 F254
glass plates. Compounds were visualized by UV, or using p-anisaldehyde, ninhydrin and
phosphomolybdic acid (PMA) as visualizing agent. E. Merck silica gel 60 (0.040–0.063 mm
particle sizes) was used for flash chromatography.
Instrumentation. Melting points were recorded on a Yanaco melting point apparatus.
Optical rotations were measured on a digital polarimeter of Japan JASCO Co. DIP-1000. [α]D
–1
2 –1
values are given in units of 10 deg cm g . Infrared (IR) spectra were recorded on Thermo
Scientific Nicolet iS5 FT-IR spectrometer. UV–Vis spectra were measured on PerkinElmer
Lamda 35 UV–Vis spectrometer. Nuclear magnetic resonance (NMR) spectra were obtained
on Varian Unity Plus-400 (400 MHz) spectrometer, and chemical shifts (δ) were recorded in
parts per million (ppm) relative to internal standards: CHCl (δ = 7.24), CDCl (δ = 77.0
3
H
3
C
ACS Paragon Plus Environment

Page 29 of 61
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
for the central line of triplet), CH OD (δ = 3.31), CD OD (δ = 49.0), (CH ) SO (δ = 2.50)
3
H
3
C
3 2
H
and (CD ) SO (δ = 39.5). The splitting patterns are reported as s (singlet), d (doublet), t
3
2
C
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(
triplet), q (quartet), m (multiplet), br (broad) and dd (doublet of doublets). Coupling
constants (J) are given in hertz (Hz). ESI–HRMS spectra were recorded on a Bruker
Daltonics BioTOF III high-resolution mass spectrometer. LC–MS/MS was conducted on
Waters ACQUITY nanoUPLC system conjugated with Waters Synapt G2 HDMS (Milford,
MA, USA). High-performance liquid chromatography (HPLC) was performed on Agilent
1100 Series instrument equipped with a degasser, Quat pump and UV detector.
Protein and ligand setup. The missing LID region in ATP-free hTMPK structure was
8
modeled based on the template (Protein Data Bank entry 1E2D). Solvent molecules were
2
+
removed, retaining TMP and Mg ions. All hydrogen atoms were added, and then the protein
3
0
structure was subjected to energy minimization using the NAMD program with the
31
CHARMM27 force field. The optimized structures and partial charges for the ligands were
3
2
obtained using the Gaussian 09 program with the 6-31G(d,p) basis set.
Docking. All docking runs were performed with version 1.1.2 of the AutoDock Vina
1
5
program. MGL Tools were used to generate the structure format (PDBQT) for Autodock
Vina, and nonpolar hydrogens were removed. First, in order to screen for the best binding
3
sites, the ligand was docked against TMPK with a large grid box 50 × 50 × 50 Å to include
2
+
the ligand and protein for a global search. Mg ions were removed. Protein was fixed and the
single bonds of ligands were flexible during the process. This approach obtains an evaluation
of scoring function during the docking process, so that as many conformations as possible can
be detected. The minimum scoring value indicates the most likely conformation. We
compared and selected the binding sites and docking affinities that have the highest ranking
ACS Paragon Plus Environment