Discovery of a potent and highly selective transforming growth factor β receptor-associated kinase 1 (TAK1) inhibitor by structure based drug design (SBDD)

Article abstract of DOI:10.1016/j.bmc.2016.07.006

A novel thienopyrimidinone analog was discovered as a potent and highly selective TAK1 inhibitor using the SBDD approach. TAK1 plays a key role in inflammatory and immune signaling, so TAK1 is considered to be an attractive molecular target for the treatm

Full text of DOI:10.1016/j.bmc.2016.07.006

Accepted Manuscript
Discovery of a potent and highly selective transforming growth factor β recep-
tor-associated kinase 1 (TAK1) inhibitor by structure based drug design (SBDD)
Terushige Muraoka, Mitsuaki Ide, Kenji Morikami, Machiko Irie, Mitsuaki
Nakamura, Takaaki Miura, Takayuki Kamikawa, Masamichi Nishihara,
Hirotaka Kashiwagi
PII:
S0968-0896(16)30498-9
DOI:
http://dx.doi.org/10.1016/j.bmc.2016.07.006
BMC 13119
Reference:
Bioorganic & Medicinal Chemistry
To appear in:
Received Date:
Revised Date:
Accepted Date:
27 May 2016
4 July 2016
5 July 2016
Please cite this article as: Muraoka, T., Ide, M., Morikami, K., Irie, M., Nakamura, M., Miura, T., Kamikawa, T.,
Nishihara, M., Kashiwagi, H., Discovery of a potent and highly selective transforming growth factor β receptor-
associated kinase 1 (TAK1) inhibitor by structure based drug design (SBDD), Bioorganic & Medicinal Chemistry
(2016), doi: http://dx.doi.org/10.1016/j.bmc.2016.07.006
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Discovery of a potent and highly selective transforming growth factor β receptor-associated kinase 1
(TAK1) inhibitor by structure based drug design (SBDD).
Terushige Muraoka a*, Mitsuaki Ide a, Kenji Morikami a, Machiko Irie b, Mitsuaki Nakamura b, Takaaki Miura b,
Takayuki Kamikawa b, Masamichi Nishihara a and Hirotaka Kashiwagia
a Research Division, Chugai Pharmaceutical Co,. Ltd., 1-135 Komakado, Gotemba, Shizuoka 412-8513, Japan
b Research Division, Chugai Pharmaceutical Co., Ltd., 200 Kajiwara, Kamakura, Kanagawa 247-8530, Japan
*Corresponding author. Tel.: +81 0550 87 6742; fax: +81 0550 87 5326
E-mail address: muraokatrs@chugai-pharm.co.jp (T. Muraoka).
Key words: TAK1, SBDD, inhibitor, kinase selectivity
Abstract
A novel thienopyrimidinone analog was discovered as a potent and highly selective TAK1 inhibitor using the
SBDD approach. TAK1 plays a key role in inflammatory and immune signaling, so TAK1 is considered to be
an attractive molecular target for the treatment of human diseases (inflammatory disease, cancer, etc.). After the
hit compound had been obtained, our modifications successfully increased TAK1 inhibitory activity and
solubility, but metabolic stability was still unsatisfactory. To improve metabolic stability, we conducted
metabolic identification. Although the obtained metabolite was fortunately a potent TAK1 inhibitor, its kinase
selectivity was low. Subsequently, to achieve high kinase selectivity, we used SBDD to follow two strategies:
one targeting unique amino acid residues in TAK1, especially the combination of Ser111 and Asn114; the other
decreasing the interaction with Tyr106 at the hinge position in TAK1. As expected, our designed compound
showed an excellent kinase selectivity profile in both an in-house and a commercially available panel assay of
over 420 kinases and also retained its potent TAK1 inhibitory activity (TAK1 IC50=11 nM).
1. Introduction
Transforming growth factor β receptor-associated kinase 1 (TAK1) is a serine/threonine kinase that plays a
key role in inflammatory and immune signaling. TAK1 was first discovered by Matsumoto et al. as a member
of the mitogen-activated kinase kinase kinase (MAP3K) family activated by transforming growth factor-β
(TGF-β),1 and since then has been elucidated as a mediator of multiple cytokines such as TNF-α, IL-1, LPS,
etc.2-4 TAK1 exerts its activity only when it forms complexes with its activator proteins TAB1 and TAB2 or
TAB3.5-7 TAB1 binds to TAK1 at the N-terminal domain, while either TAB2 or TAB3 binds to TAK1 at the
C-terminal domain. TAK1 phosphorylates downstream kinases, such as the mitogen-activated kinase kinase
(MKK) family (MKK3/4/6/7), p38, c-Jun N-terminal kinase (JNK), and the IκB kinase (IKK) family (IKKα,
IKKβ, and NEMO), to activate transcription factors, such as activation protein-1 (AP-1) and nuclear factor-κB
(NF-κB), which regulate stress response, immunity, inflammation, and cancer signaling. Indeed, p38 and IKK

have been developed as targets for rheumatoid arthritis and airway inflammation,8 and systemic administration
of TAK1 siRNA particles has been shown to alleviate inflammation in a mouse model of immune-mediated
inflammatory disorder.9 In summary, TAK1 is considered to be an attractive molecular target for the treatment
of human diseases (inflammatory disease, cancer, etc.).
It is important that compounds designed as kinase inhibitors have a good kinase selectivity profile because
inhibiting particular kinases carries the risk of adverse effects. For example, inhibiting KIT and KDR is related
to bone marrow suppression and hypertension, respectively.10, 11 Additionally, although FLT3 is an attractive
target for acute myelogenous leukemia and acute lymphocytic leukemia, it plays a key role in the normal
development of stem cells and the immune system in hematopoiesis;12 therefore, we considered that FLT3
inhibition should be avoided, especially in chronic diseases. Of over 500 kinases in human13 that maintain the
functions of cells, it is not completely clear which kinase inhibition causes an adverse event. A further reason
why compounds need to have high kinase selectivity is to elucidate the mode of action more easily. Even if a
multi kinase inhibitor achieves the desired phenotype, we are still not able to tell whether its desired effect is a
consequence of target inhibition or not.
Several TAK1 inhibitors have been reported previously (Figure 1)14. 5Z-7-oxazeaenol 115, 16 and epoxyquinol
B 217, 18, the natural products known as irreversible inhibitors, are unselective kinase inhibitors that nevertheless
have potent TAK1 inhibitory activity. Thiophene carboxamide 319 and oxindole 420 also have potent TAK1
inhibitory activity, but their selectivity data has not been reported. Although the TAK1 IC50 of
imidazopyridazine 521 is 8.0 nM, a kinase panel showed that it had IC50 < 100 nM to 10/30 kinases other than
TAK1, so we consider its TAK1 selectivity to be low. The recently discovered TAK1 inhibitor,
7-aminofuro[2,3-c]pyridine 622, 23 (OSI Pharmaceuticals), showed >50% inhibition for 4/192 kinases (FLT3,
HGK, MAP4K5, and MINK1) at fifty times the concentration of TAK1 Ki, so can be considered relatively
selective, but as explained above its FLT3 inhibitory activity would be undesirable. Whereas compounds 1~6,
which bind TAK1 in the DFG-in conformation, are known as type I inhibitors, Pfizer has reported the first type
II inhibitor, pyrazole urea 724, which binds TAK1 in the DFG-out conformation. According to the preliminary
study, TAK1 inhibitory activity and selectivity of compound 7 are not sufficient.
Figure 1. Structure of previously reported TAK1 inhibitors (1-7)

Here, we report the discovery of novel potent and selective TAK1 inhibitor 12 that possesses a
thienopyrimidinone moiety as a scaffold (Figure 2). Our starting compound 8 was obtained from two focused
libraries selected from in-house compounds; the first library was selected from inhibitors against five kinases
that either had similar amino acid sequences or for which cross activity against TAK1 kinase was strongly
predicted by the existing patent information, and the second library included kinase inhibitors with excellent
kinase selectivity profile (fewer than five kinases with Kd < 1 µM) as characterized by KinomeScanTM
(DiscoverRX, an in vitro binding panel assay system). We assayed their binding activity to TAK1 with SPR
(manuscript in preparation) and refer to the use of these focused libraries as the “chemogenomics approach”.
Compound 8, which showed the expected good TAK1 selectivity profile but had extremely low solubility, was
then modified to increase TAK1 inhibitory activity and solubility, which resulted in compounds 11c and 11i
with potent TAK1 inhibitory activity. These compounds were then further modified to improve their kinase
selectivity profile by fully utilizing the information from the X-ray structures of the inhibitor-TAK1 complexes.
As a result, we were able to discover compound 12 with both potent TAK1 inhibitory activity (TAK1 IC50=11
nM) and high kinase selectivity. The in-house panel data of 38 kinases, including FLT3, showed no significant
inhibition (IC50 > 10µM) other than that against TAK1. More significantly, in a KinomeScanTM panel of 442
kinases, 12 showed a 50% binding inhibition larger than 1.0 µM against as few as four kinases, one of which
was TAK1.
Figure 2. Structure of Hit compound (8) and our designed TAK1 inhibitors.
2. Chemistry
The synthesis of thienopyimidine 16 and thienopyrimidinone 19 is shown in Scheme 125. The common
intermediate 14 of both 16 and 19 was synthesized from commercially available thiophene 13 in moderate
yield (2 steps 72%). The hydrogenation of 14 with palladium hydroxide generated 15 in 94% yield.

Subsequently, dibromination of 15 at the benzylic position obtained the aryl aldehyde, which was then
subjected to oxidation reaction to give the desired thienopyrimidine 16 (3 steps 58%). As for the synthesis of
thienopyrimidinone 19, compound 17 was obtained first by dibromination of 14. After the generation of
trimethoxy compound by treating 17 with methanol, aldehyde 18 was obtained under acidic condition. Finally,
the oxidation of the aldehyde gave the desired thienopyrimidinone 19 in good yield (3 steps 75%).
Scheme 1. Synthesis of thienopyrimidine 16 and thienopyrimidinone 19. Reagents and conditions: (a) Formamide, 150 °C; (b)
POCl3, 100 °C, 2 steps 72%; (c) H2 gas (1.0 atm), 20% Pd(OH)2/C, Et3N, EtOAc/i-PrOH = 5/1, room temp. 94%; (d) NBS,
benzoyl peroxide, CCl4, reflux; (e) Pyridine, H2O, 100 °C; (f) NaClO2, sulfamic acid, acetone, H2O, 40 °C, 3 steps 58%; (g) NBS,
AIBN, CCl4, reflux; (h) MeOH, 70 °C and then 5N HCl aq., THF, 60 °C.; (i) NaClO2, sulfamic acid, acetone, H2O, room temp., 3
steps 71%.
The synthesis of thienopyrimidine analogs (9a-9c) by amidation is shown in Scheme 2. Aniline parts
(20a-20c) were prepared using a commercially available substituted benzene 21 or 22 in good yield.
Thienopyrimidinone analogs 10, 11a-l, and 12 were synthesized by a similar method as the thienopyrimidine
analogs 9a-c (Scheme 3). The preparation method of aniline parts is shown in Supplementary Scheme 1.
Scheme 2. Synthesis of thienopyrimidine analog. Reagents and conditions : (a) HATU, DIPEA, 20a-c, acetonitrile; (b) amine,
HATU, DIPEA, DCM; (C) H2, 10% Pd/C, ethanol; (d) N-methyl piperazine, Et3N, DMA; (e) H2, 10% Pd/C, methanol.

Scheme 3. Synthesis of thienopyrimidinone analogs. : (a) HATU, DIPEA, aniline, acetonitrile.
3. Results and discussion
We found the hit compound 8 by SPR screening but its low solubility made it impossible to evaluate TAK1
inhibitory activity at more than 1.0 µM (37% inhibition at 1.0 µM). Indeed, its solubility measured by fasted
state simulated intestinal fluid (FaSSIF) or phosphate buffer (PPB) at pH 6.5 was less than 3.0 µg/mL because
of its planar structure26.
First, the quinoline ring of hit compound 8 was substituted for the benzene ring bearing a hydrophilic group,
to improve the solubility. The dissolution of the resultant compounds 9a-c was much improved both in FaSSIF
and in PPB, and 9b also exhibited TAK1 inhibitory activity three times more potent than the hit compound 8,
but metabolic stability, especially in mouse microsomes, was low compared to human (Table 1).

Table 1. TAK1 inhibitory activity and physicochemical parameters of HTS hit compound (8) and thienopyrimidine analogs
(9a-9c)
FaSSIF(PPB)
CLint(µL/min/mg)
compound
R1
Kinase IC50 (µM)
(µg/mL)
human
mouse
-
37% inhibition
-
-
Hit compound 8
< 3 (< 3)
312 (293)
420 (377)
@ 1µM
1.6
0.26
1.3
26
21
-
9a
9b
80
11
-
9c
351 (353)
In order to identify the metabolically labile site of 9b, its metabolites were structurally analyzed, particularly
one main metabolite observed in both human and mouse liver microsomes, metabolite 10, which possessed a
potent TAK1 inhibitory activity (IC50=13 nM) (Figure 3(a)). Because we discovered this metabolite was
generated in liver cytosol, the main metabolite 10 was isolated using 9b in mouse liver cytosol to determine the
structure by LC-MS and NMR, which showed that an oxidation reaction at 4 position of the thienopyrimidine
ring of 9b had occurred. Metabolite 10 was found to be a potent TAK1 inhibitor (IC50=13 nM), so we used
X-ray analysis on an opportunely obtained crystal structure of metabolite 10 in TAK1 to understand this
increased TAK1 inhibitory activity.
The X-ray crystal structure showed 10 was binding to TAK1 in the DFG-in conformation known as a type I
inhibitor (Figure 3(b)). Focusing on how differences between 9b and 10 increased TAK1 inhibitory activity, we
found that the critical difference was a carbonyl group of thienopyrimidinone located near the DFG loop and
the gatekeeper of TAK1. The X-ray information revealed that the distance from this carbonyl group to crystal
water was 2.9 Å and the distance from the crystal water to the backbone amide NH of Asp175 (DFG) was 2.9
Å. In other words, this carbonyl group of thienopyrimidinone has tight interactions with the backbone amide
NH of Asp175 by hydrogen bonding via a crystal water. In addition, we argue that this carbonyl group also has
a tight interaction with the side chain of Glu77, though because this is not clear in the X-ray crystal structure,
we confirmed that some derivatives with a thienopyrimidinone ring like compound 10 interacted with the side
chain of Glu77 via this crystal water (PDB code: 5JGA and 5JGD). As a result of these tight interactions,

desirable effects on the potent inhibitory TAK1 activity of 10 were obtained in comparison to 9b. Next, we
elucidated another interaction between 10 and TAK1, with the distance from amide NH at 3-position of the
thienopyrimidinone ring to the side chain of DFG Asp175 of 2.6 Å, and the distance from the S atom at
5-position of thienopyrimidinone to the side chain of gatekeeper Met104 of 4.0 Å, which meant that
thienopyrimidinone of 10 had a tight interaction with DFG Asp175 by hydrogen bonding and a weak
hydrophobic interaction with Met104. Subsequently, in the hinge position, the distance from an amide carbonyl
group at 7-position of the thienopyrimidinone part of 10 to Ala107 was 2.7 Å, which created a hydrogen
bonding interaction with TAK1 in this region. Finally, at N-methyl piperazine moiety located in the solvent
region, no tight interaction was found. These interactions between 10 and TAK1, especially those between the
carbonyl group of the thienopyrimidinone ring and TAK1 via crystal water, led to potent TAK1 inhibitory
activity (IC50=13 nM).
Figure 3. (a) Metabolic reaction of 9b in liver cytosol. (b) X-ray crystal structure of 10 (metabolite of 9b) in TAK1 (PDB
code: 5JGB). 10 is shown as a ball and stick model colored by element type (C in magenta, O in red, N in blue, S in yellow). Amino
acid residues of TAK1 that interact with compound 10 are shown as stick models colored by element type with the same colors,
except for C in light green. A representative crystallographic water is depicted as a red sphere. The key interactions are depicted as
dotted lines.
Subsequently, we explored the SAR of R2 group (Table 2), which the X-ray crystal structure showed was
located in the vicinity of hydrophobic amino acids Val42, Gly43, Gly45, Val50, Gly110, Pro160, and Leu163.
Thus, we defined this region as hydrophobic and introduced a hydrophobic R2 group in this region to enhance
the interaction with TAK1. As for the linker connecting the benzene ring and R2 groups, three different types
(C-linker, O-linker and N-linker) were prepared to change the direction of the R2 group, and halogen
derivatives were also evaluated. In each case, the volume of the R2 group was calculated while ignoring the
hydrogen atoms. The tendency obtained with each linker type was very similar. As expected, TAK1 inhibitory

activity increased as the volume of the hydrophobic R2 group increased; however, reduced TAK1 inhibitory
activity was observed when the volume of the R2 group was beyond the local maximum value of each linker
type (Figure 4). Regarding the TAK1 inhibitory activity of C-linker derivatives 11a-11d, 11c had the most
potent (2.3 nM); 11a, which had insufficient occupancy in this hydrophobic region, showed the weakest (41
nM); while 11d, which possessed the largest substituent, cyclohexane, was slightly weaker (6.7 nM) than 11c.
Halogen derivatives (11e-11g) and O-linker derivatives (11h-11j) had the same tendency as C-linker
derivatives; chlorine derivative 11f and iPrO derivative 11i group were the most potent (5.8 nM and 3.0 nM,
respectively); fluorine derivative 11e (220 nM) and OMe derivative 11h (15 nM) were the weakest, with
insufficient volume of R2 group; but 11g (7.3 nM) and 11j (5.2 nM), which possessed the largest substituent (a
bromine and hexafluoroisopropoxy group), had slightly weaker activity than 11f and 11i, respectively.
Regarding TAK1 inhibition of N-linker derivatives, 11k with a dimethyl amino group was 5.0 nM, while 11l
with an ethyl methyl amino group was 9.3 nM; therefore, we estimated that the local maximum value of
N-linker types would be below 65 Å3 of volume. We considered that the difference in local maximum value of
TAK1 inhibitory activity shown by each linker type depended on the difference in the direction of the benzene
ring in R2. Having discovered potent TAK1 inhibitors 11c and 11i (IC50=2.3 nM and 3.0 nM, respectively), we
evaluated the kinase selectivity of these compounds, but it was unfortunately low. The result of the in-house
panel assay of 38 kinases showed that 11c also had inhibitory activity below 1.0 µM against 14 kinases other
than TAK1. In particular, 11c had potent inhibitory activity against FLT3 (IC50=5.4 nM), which we felt should
be avoided, and 11i also had the same tendency as 11c. By exploiting the hydrophobic pocket, we had
succeeded in finding TAK1 inhibitors as potent as 11c and 11i, but their kinase selectivity profile was rather
compromised, presumably because the ATP sugar part occupied this pocket in TAK1 and the amino acid
residues around this pocket are common to many kinases

Table 2. SAR table for TAK1 activity and calculated volume of R2 group.
a calculated volume of side chain moiety (R2 group) extracted from whole molecule.
Kinase
Volume
Compound
R2
(Å3)a
IC50 (nM)
11e
11f
11g
220
5.8
7.3
12
25
27
Kinase
Volume
Compound
R2
11h
11i
15
32
64
(Å3)a
IC50 (nM)
11a
41
27
56
3.0
11b
3.5
11j
5.2
95
11c
11d
2.3
6.7
66
87
11k
11l
5.0
9.3
52
65

Figure 4. Relationship between TAK1 activity and volume of R2 group in thienopyrimidinone analogs (11a-11l). (magenta: C-linker
derivative, yellow: halogen derivative, light green: O-linker derivative, cyan: N-linker derivative)
a calculated volume of side chain moiety (R2 group) extracted from whole molecule.
To improve the kinase selectivity of 11c or 11i, we adopted two strategies that utilized 3D structure
information. The first one was to target specific amino acid residues in TAK1, especially the combination of
Ser111 and Asn114, which was revealed to be very unique from a search of an in-house database constructed
from literature on 490 kinases13 (Figure 5 (b)). The second strategy was to reduce the interaction between the
amide moiety located at R4 position in the substituted benzene and Tyr106 at the hinge position in TAK1. For
the first strategy, we made use of previous reports22, 23 by OSI Pharmaceuticals in which Ser111 was targeted to
obtain selectivity to TAK1 over Aurora B. Because Ser111 in TAK1 corresponds in Aurora B to the larger
Glu161, they suggested that clashing with the Glu161 in Aurora B meant that their compound 6 was more
selective to TAK1 than to Aurora B, but moderately selective to FLT3 (at fifty times the concentration of TAK1
Ki, 6 showed > 50% FLT3 inhibition). We found that out of 490 kinases, position 111 was occupied by Ser in
115 kinases (as in TAK1), by Glu in 91 kinases (as in Aurora B), and by Asp in 161 kinases (as in FLT3). We
considered that the difference between Asp and Glu made compound 6 more selective to Aurora than to FLT3,
because the size of Asp is smaller than Glu, so compound 6 bumped less against FLT3. To obtain high TAK1
kinase selectivity, we focused on the combination of Ser111 and Asn114 in TAK1. As mentioned above, Ser is a
fairly common amino acid residue, but only 20 kinases had Asn corresponding to the 114 position, as TAK1
does, and possessing both Ser and Asn at these positions is very rare and unique; in fact, only one kinase other

than TAK1, integrin-linked kinase (IIK), has this combination. From the X-ray information of compound 11c in
TAK1, the distance from the carbonyl group of R4 group in the solvent region to Ser111 and Asn114 was 6.7 Å
and 4.7 Å, respectively, which indicated that the compound had insufficient interaction against the area
surrounding these TAK1 unique amino acids (Figure 5 (a)). We supposed that the desirable interaction with
TAK1 at Ser111 and Asn114 would improve kinase selectivity.
For the second strategy, we did a similar search of 490 kinases for amino acid residues corresponding to
Tyr106 in TAK1, and found that 198 have Tyr in this position, and additionally many kinases (304) have
aromatic amino acid residues there. The X-ray structure of 11c in TAK1 showed that the distance from the
N-methyl piperazine ring at R4 group to Tyr106 is 3.4 Å, which indicated a CH-π interaction27 between 11c and
TAK1. We assumed that the same interaction between 11c and other kinases exists, and that reduction of this
interaction would improve kinase selectivity.
We designed compound 12 to adopt a simple but efficient plan that was able to achieve these two different
strategies simultaneously. Specifically, N-methyl piperazine amide moiety at R4 position of 11c was shifted to
R3 position, which we expected to make the distance from the amide moiety of 11c to Ser111 and Asn114
closer, while making that of 11c to Tyr106 further. To evaluate our design of compound 12, we superimposed
the crystal structure of FLT3 (PDB code:1RJB)28 to that of TAK1 (Figure 5 (C) right) and found a critical
difference between the structure of TAK1 and FLT3; namely, that near the amide carbonyl group of 12 at R3
position, Ser111 was located in TAK1, but Asp698 in FLT3. With TAK1, the amide carbonyl interacted with
Ser111, whereas in FLT3, this amide carbonyl had repulsive coulomb against Asp698. We supposed this
difference could give 12 TAK1 selectivity against FLT3.
As expected, 12 showed an excellent kinase selectivity profile in both in-house and KinomeScanTM panel
assays and possessed potent TAK1 inhibitory activity (IC50=11 nM). In the in-house panel data of 38 kinases,
we could not see potent inhibitory activity (IC50 > 10 µM) except against TAK1. Particularly, the IC50 of FLT3
was 35000 nM, which indicated that compound 12 had about 3200 times more potent inhibitory activity on
TAK1 than on FLT3. In the 442-kinase panel data, 12 showed inhibition against only 3 kinases other than
TAK1 at 1.0 µM. The other 3 kinases are MAP4K4, PIP5K1C, and PIP5K2C. MAP4K4 is an upstream kinase
of TAK1, and the other two kinases are lipid kinases. To find out why 12 had a high TAK1 selectivity profile,
the X-ray structure of TAK1 and 12 complex was solved (Figure 5 (C) left). The binding mode of 11c and 12
was very similar basically, but three differences existed that would afford the high TAK1 selectivity profile of
12. The first one was acquiring the interaction with Ser111 and Asn114, which was a rare and unique
combination of amino acid residues in TAK1. As expected, the distances from the carbonyl group at R3 position
of 12 to Ser 111 and Asn114 were 3.4 Å and 4.2 Å, respectively, which is closer than those in 11c. The second
one was that compound 12 had repulsive coulomb interactions against kinases with Asp or Glu corresponding
to Ser111 in TAK1. From our analysis, 252 out of 490 kinases had Asp or Glu corresponding to Ser111 in
TAK1, and these kinases were considered to possess a repulsive coulomb interaction against compound 12,
which resulted in TAK1 selectivity. The third one was reducing the interaction between the inhibitor and

Tyr106 in TAK1 at the hinge region. The distance from the benzene ring of 12 to Tyr106 in TAK1 is 4.2 Å. In
contrast, the distance form N-methyl piperazine ring of 11c to Tyr106 is 3.4 Å. 304/490 kinases have an
aromatic residue corresponding to Tyr106 in TAK1. Decreasing this interaction reduced the activity of the
above-mentioned 304 kinases, including TAK1. Even though TAK1 inhibitory activity was also decreased, we
considered that TAK1 selectivity was improved. According to these factors, we were able to discover
compound 12 which was a potent TAK1 inhibitor with excellent selectivity.
Figure 5. (a) X-ray crystal structure of 11c in TAK1(PDB code:5JGA). Compound 11c is shown as a ball and stick model colored
by element type (C in magenta, O in red, N in blue, S in yellow). Amino acid residues of TAK1 that interact with compound 11c are
shown in the same colors except for C in light green. The key hydrogen bonding interactions are depicted as dotted lines (b) Amino
acid residues corresponding to Tyr106, Ser 111, and Asn114 of TAK1 in a 490 kinase in-house database. (c) 12 in the crystal
structure of TAK1 (left, PDB code:5JGD) and in that FLT3 (right). Compound 12 is shown as a ball and stick model colored by
element type (C in magenta, O in red, N in blue, S in yellow). Amino acid residues of TAK1 and FLT3 that interact with compound
12 are shown in the same colors except for C, which is light green in TAK1 and cyan in FLT3. The key hydrogen bonding
interactions are depicted as dotted lines.
4. Conclusion
In summary, we have discovered a potent inhibitor that is highly selective to TAK1 by using the SBDD
approach to develop a novel series of thienopyrimidinones. After hit compound 8 was obtained by a

chemogenomics approach, it was modified to give 9b with increased TAK1 inhibitory activity and solubility,
but its metabolic stability was low. Metabolites identified from 9b included 10 with potent TAK1 inhibition.
X-ray information revealed that the carboxyl group in thienopyrimidinone played a key role in acquiring the
potent TAK1 inhibitory activity. Next, we were able to obtain a more potent TAK1 inhibitor 11c (IC50=2.3 nM)
by adjusting the volume of R2 group, but TAK1 selectivity was low. To improve this kinase selectivity, we
conducted two strategies utilizing 3D structure information: one targeting unique amino acid residues in TAK1,
especially the combination of Ser111 and Asn114; the other decreasing the interaction between the amide
moiety at R4 group in the solvent region and Tyr106 at the hinge position in TAK1, which has an aromatic side
chain that is highly conserved between protein kinases. As expected, 12 showed an excellent TAK1 selectivity
profile in both an in-house and a KinomeScanTM panel assay and possessed potent TAK1 inhibitory activity
(IC50=11 nM). Particularly, the in-house 38-kinase panel data did not reveal potent inhibitory activity
(IC50 < 10 µM) of kinases other than TAK1.
We could obtain the potent and highly selective TAK1 inhibitor 12, but cell activity was weak (IC50 > 5 µM,
SW982 cells :TNF-α-induced IL-6 production) because of low permeability (Caco-2 Papp < 1.0x10-6 cm/sec).
So, further optimization of this series to improve permeability and other physicochemical property is ongoing.
5. Experimental
5.1. Chemistry: general
1
Purchased reagents and solvents were used without further purification unless otherwise noted. H and 13C
spectra were recorded on VARIAN 400-MR or Bruker AVANCE 2 or 3 and chemical shifts are expressed in
parts per million (ppm) from trimethylsilane (TMS) as an internal standard. The spectrum patterns are
described as follows: s = singlet, d = doublet, t = triplet, q = quartet, br = broad, m = multiplet. LC-MS were
measured by a Waters ACQUITY SQD electrospray ionization (ESI) system. High resolution mass spectra
(HRMS) were measured with a Waters I-Class/Xevo G2S TOF MS spectrometer using an ESI source coupled
to a Waters HPLC system operating in reversed phase with an Ascentis Express C18 (2.7 µm, 2.1 mm x 50
mm) column or no column. Flash column chromatography was performed with Purif-pack® or Biotage SNAP®
cartridges. NH-silica gel chromatography and preparative TLC were conducted with FUJI SILYSIA
NH-DM1020 and Merck silica gel 60 PF254, respectively. Reversed column chromatography was conducted
with an ODS column, Wakosil® 25C18.
5.1.1. 4-Chloro-7-methylthieno[3,2-d]pyrimidine (14)
Commercially available methyl 3-amino-4-methylthiophene-2-carboxylate 13 (16.5 g, 96.0 mmol) was
dissolved in formamide (100 mL) and stirred at 150 °C under nitrogen atmosphere. After the starting material
consumption had been confirmed by LC-MS, the reaction mixture was cooled to room temperature and then

H2O (500 mL) was added to the mixture. The suspension was filtered to collect a white solid that was dried by
an oil pump to give 7-methylthieno[3,2-d]pyrimidin-4(3H)-one (12.2 g). The solid was used for the next
reaction without further purification.
The aforementioned 7-methylthieno[3,2-d]pyrimidin-4(3H)-one (1.85 g) was treated with POCl3 (8.80 mL)
and stirred at 100 °C for 6h. The reaction mixture was cooled to 0 °C, poured into ice-water/ethyl acetate, and
extracted with ethyl acetate twice. The organic layer was washed with brine, dried over Na2SO4, and filtered.
The organic solvent was concentrated under reduced pressure to give a crude residue. Water (200 mL) was
added to the residue and the mixture cooled at 5 °C in the refrigerator for a day. The suspension was filtered to
1
afford 14 (1.90 g, 2steps 72%) as a white solid. H-NMR (400 MHz, DMSO-D6) δ: 9.07 (1H, s), 8.25 (1H, d, J
= 1.2 Hz), 2.45 (3H, d, J = 1.2 Hz). 13C-NMR (100 MHz, DMSO-D6) δ: 160.7, 153.9, 153.6, 133.9, 133.2,
130.0, 12.5. HRMS (ESI-TOF) m/z: [M+H]+ Calcd for C7H6ClN2S 184.9940; Found 184.9937.
5.1.2. 7-Methylthieno[3,2-d]pyrimidine (15)
A solution of ethyl acetate/2-propanol = 5/1 (v/v, 200 mL) was added to Pd(OH)2 (4.83 g, 20% on carbon,
wetted with ca.50% water) under nitrogen atmosphere and filled with hydrogen. A solution of 14 (9.65 g, 51.7
mmol) in ethyl acetate/2-propanol = 5/1 (v/v, 50.0 mL) and Et3N (16.6 mL, 119 mmol) was added to the
suspension and filled with hydrogen again. The suspension was stirred at room temperature for 20 h. The
reaction mixture was filtered with Celite® and concentrated. After the addition of ethyl acetate and saturated
NaHCO3 aq. to the residue, the ethyl acetate extraction was executed twice. The combined organic layer was
washed with brine, dried over Na2SO4, and filtered. The organic solution was concentrated under reduced
pressure to afford 15 (7.30 g, 94%). 1H-NMR (400 MHz, DMSO-D6) δ: 9.50 (1H, s), 9.19 (1H, s), 8.15 (1H, d,
J = 1.2 Hz), 2.44 (3H, d, J = 1.2 Hz). 13C-NMR (100 MHz, DMSO-D6) δ: 159.0, 153.8, 152.4, 133.1, 132.0,
131.3, 12.1. HRMS (ESI-TOF) m/z: [M+H]+ Calcd for C7H7N2S 151.0330; Found 151.0332.
5.1.3. Thieno[3,2-d]pyrimidine-7-carboxylic acid (16)
A solution of 15 (1.00 g, 6.66 mmol), NBS (2.97 g, 16.7 mmol) and benzoyl peroxide (322 mg, 1.33 mmol)
in CCl4 (67.0 mL) was stirred at reflux for 2.5 h. The reaction mixture was cooled to room temperature and
filtered. The organic solvent was evaporated to give 7-(dibromomethyl)thieno[3,2-d]pyrimidine as a crude
residue. The residue was used for the next reaction without further purification.
To the aforementioned residue was added H2O (40.0 mL) and pyridine (4.00 mL) and the mixture was stirred
at 100 °C for 1.5 h. The reaction mixture was diluted with ethyl acetate and extracted. The organic layer was
washed with brine, dried over Na2SO4 and concentrated. The obtained white solid was washed with
hexane/ethyl acetate = 4/1 (v/v) to afford thieno[3,2-d]pyrimidine-7-carbaldehyde (796 mg) as a crude residue.
The residue was used for the next reaction without further purification.
To a solution of the aforementioned residue (796 mg) in acetone/H2O = 2/1 (v/v, 99.0 mL) was added
NaClO2 (658 mg, 7.28 mmol) and sulfamic acid (1.17 g, 12.1 mmol) at 0 °C and then stirred at 40 °C for 2 h.

The reaction mixture was evaporated to give a white solid. The solid was washed with H2O and ethyl acetate to
1
afford 16 (690 mg, 3 steps 58%). H-NMR (400 MHz, DMSO-D6) δ: 12.98 (1H, br s), 9.63 (1H, s), 9.28 (1H,
s), 9.18 (1H, s). 13C-NMR (100 MHz, DMSO-D6) δ: 162.0, 156.8, 154.6, 153.4, 147.0, 132.0, 127.3. HRMS
(ESI-TOF) m/z: [M+H]+ Calcd for C7H5N2O2S 181.0071; Found 181.0069.
5.1.4. 4-Chloro-7-(dibromomethyl)thieno[3,2-d]pyrimidine (17)
To a solution of 14 (4.00 g, 21.4 mmol) in CCl4 (43.0 mL) was added NBS (9.50 g, 53.5 mmol) and AIBN
(703 mg, 4.28 mmol) at room temperature. The reaction mixture was stirred at reflux for 4 h under nitrogen
atmosphere. After the reaction mixture was filtered and evaporated, the obtained residue was purified by flash
column chromatography (hexane/ethyl acetate) to afford 17 (6.92 g) as a semi-pure product. The residue was
used for the next reaction without further purification.
5.1.5. 4-Oxo-3,4-dihydrothieno[3,2-d]pyrimidine-7-carbaldehyde (18)
The above semi-pure 17 (6.92 g) was treated with methanol (200 mL) and stirred at 70 °C for 24 h to
generate 7-(dimethoxymethyl)-4-methoxythieno[3,2-d]pyrimidine. The reaction mixture was evaporated and
the residue was diluted with THF (100 mL). To the solution was added 5N HCl aq. (100 mL) and the mixture
was stirred at 60 °C for 3 h. After the reaction mixture was evaporated, the obtained solid was collected by
filtration and dried under reduced pressure to afford 18 as a crude product. The crude product was used for
the next reaction without further purification.
5.1.6. 4-Oxo-3,4-dihydro-thieno[3,2-d]pyrimidine-7-carboxylic acid (19)
To a solution of the above crude product 18 in acetone (200 mL) and H2O (100 mL) was added NaClO2
(2.90 g, 32.1 mmol) and sulfamic acid (5.20 g, 53.5 mmol). The reaction mixture was stirred at room
temperature for 2 h. After the reaction mixture was evaporated, the obtained solid was washed with H2O and
1
acetone and dried under reduced pressure to afford 19 (2.96 g, 3 steps 71%) as a white solid. H-NMR (400
MHz, DMSO-D6) δ: 12.90-12.79 (2H, br m), 8.85 (1H, s), 8.27 (1H, s). 13C-NMR (100 MHz, DMSO-D6) δ:
162.1, 157.3, 154.6, 147.5, 142.3, 128.5, 124.4. HRMS (ESI-TOF) m/z: [M-H]- Calcd for C7H3N2O3S
194.9865; Found 194.9865.
5.1.7. N-(2-methoxy-4-((3-(4-methylpiperazin-1-yl)propyl)carbamoyl)phenyl)thieno[3,2-d]pyrimidine-7-
carboxamide (9b)
To a solution of 4-amino-3-methoxy-N-(3-(4-methylpiperazin-1-yl)propyl)benzamide 20b (52.0 mg, 0.170
mmol) and thieno[3,2-d]pyrimidine-7-carboxylic acid 16 (31.0 mg, 0.170 mmol) in acetonitrile (3.50 mL) was
added HATU (129 mg, 0.340 mmol) and DIPEA (59.0 µL, 0.340 mmol) and stirred at 80 °C. After confirming
the starting material consumption by LC-MS, the reaction mixture was concentrated to give a crude residue.
The crude residue was purified by column chromatography utilizing NH silica gel (DCM/methanol) to afford

9b (54.0 mg, 68%) as a yellow solid. 1H-NMR (400 MHz, CDCl3) δ: 11.91 (1H, s), 9.40 (1H, s), 9.38 (1H, s),
9.05 (1H, s), 8.69 (1H, d, J = 8.8 Hz), 8.29 (1H, t, J = 4.6 Hz), 7.64 (1H, d, J = 1.5 Hz), 7.39 (1H, dd, J = 8.3,
1.5 Hz), 4.11 (3H, s), 3.60-3.59 (2H, m), 2.75-2.63 (10H, m), 2.33 (3H, s), 1.82-1.80 (2H, m). 13C-NMR (100
MHz, DMSO-D6) δ: 165.4, 158.4, 155.2, 154.3, 153.9, 148.1, 146.7, 132.5, 130.2, 130.0, 128.5, 119.9, 118.5,
109.8, 56.3, 55.7, 54.8, 52.7, 45.8, 38.0, 26.3. HRMS (ESI-TOF) m/z: [M+H]+ Calcd for C23H29N6O3S
469.2022; Found 469.2027.
5.1.8. N-(2-methoxy-4-(4-methylpiperazine-1-carbonyl)phenyl)thieno[3,2-d]pyrimidine-7-
carboxamide (9a)
1
Compound 9a was synthesized by a similar procedure to 9b. The yield was 36%. Yellow solid. H-NMR
(400 MHz, CDCl3) δ: 11.83 (1H, s), 9.40 (1H, s), 9.38 (1H, s), 9.05 (1H, s), 8.63 (1H, d, J = 7.8 Hz), 7.11 (1H,
d, J = 2.0 Hz), 7.06 (1H, dd, J = 8.3, 1.5 Hz), 4.06 (3H, s), 3.73-3.71 (4H, m), 2.44 (4H, br s), 2.34 (3H, s).
13C-NMR (100 MHz, DMSO-D6) δ: 168.6, 158.3, 155.2, 154.3, 153.9, 148.3, 146.7, 132.5, 131.3, 128.8, 128.5,
119.6, 118.9, 110.0, 56.4, 45.6. HRMS (ESI-TOF) m/z: [M+H]+ Calcd for C20H22N5O3S 412.1443; Found
412.1444.
5.1.9 N-(2-methoxy-4-(4-methylpiperazin-1-yl)phenyl)thieno[3,2-d]pyrimidine-7-carboxamide (9c)
1
Compound 9c was synthesized by a similar procedure to 9b. The yield was 57%. Yellow solid. H-NMR
(400 MHz, CDCl3) δ: 11.53 (1H, s), 9.38 (1H, s), 9.35 (1H, s), 9.00 (1H, s), 8.44 (1H, dd, J = 6.1, 3.1 Hz), 6.58
(2H, dd, J = 6.7, 2.4 Hz), 4.01 (3H, s), 3.23 (4H, t, J = 5.2 Hz), 2.61 (4H, t, J = 4.9 Hz), 2.37 (3H, s). 13C-NMR
(100 MHz, DMSO-D6) δ: 157.5, 155.2, 154.2, 153.8, 149.6, 148.3, 145.6, 132.5. 129.1, 120.4, 119.7, 106.7,
99.8, 56.1, 54.6, 48.4, 45.8. HRMS (ESI-TOF) m/z: [M+H]+ Calcd for C19H22N5O2S 384.1494; Found
384.1501.
5.1.10. 4-amino-3-methoxy-N-(3-(4-methylpiperazin-1-yl)propyl)benzamide (20b)
To a solution of commercially available 3-methoxy-4-nitrobenzoic acid 21 (100 mg, 0.507 mmol) and
3-(4-methylpiperazin-1-yl)propan-1-amine (159 mg, 1.01 mmol) in DCM (5.00 mL) was added HATU (386
mg, 1.01 mmol) at room temperature and stirred at the same temperature. After confirming the starting material
consumption by LC-MS, the reaction mixture was concentrated to give a crude residue. The crude residue was
purified by column chromatography utilizing NH silica gel (DCM/methanol) to afford
3-methoxy-N-(3-(4-methylpiperazin-1-yl)propyl)-4-nitrobenzamide (quant).
Ethanol (3.50 mL) was added to 10% Pd/C (76.0 mg) under nitrogen atmosphere and filled with hydrogen. A
solution of 3-methoxy-N-(3-(4-methylpiperazin-1-yl)propyl)-4-nitrobenzamide (151 mg, 0.449 mmol) in
ethanol (1.00 mL) was added to the suspension, filled with hydrogen again and then stirred at 40 °C. After the
starting material consumption had been confirmed by LC-MS, the reaction mixture was filtered with Celite®
and concentrated. The obtained residue was purified by column chromatography utilizing NH silica gel

(DCM/methanol) to afford 20b (118 mg, 86%) as a light yellow oil. 1H-NMR (400 MHz, CD3OD) δ: 7.33 (1H,
d, J = 1.5 Hz), 7.28 (1H, dd, J = 8.3, 2.0 Hz), 6.70 (1H, d, J = 8.3 Hz), 3.89 (3H, s), 3.39 (2H, t, J = 6.8 Hz),
3.00-2.28 (10H, m), 2.27 (3H, s), 1.83-1.77 (2H, m). 13C-NMR (100 MHz, CD3OD) δ: 170.4, 147.9, 142.4,
123.9, 122.1, 114.3, 110.5, 57.3, 56.1, 55.6, 53.7, 45.9, 39.4, 27.4. HRMS (ESI-TOF) m/z: [M+H]+ Calcd for
C16H27N4O2 307.2134; Found 307.2146.
5.1.11. (4-amino-3-methoxyphenyl)(4-methylpiperazin-1-yl)methanone (20a)
Compound 20a was synthesized by a similar procedure to 20b. The yield was 80% (2 steps). Light yellow
oil. 1H-NMR (400 MHz, CDCl3) δ: 6.94 (1H, d, J = 1.5 Hz), 6.86 (1H, dd, J = 8.1, 1.7 Hz), 6.65 (1H, d, J = 8.3
Hz), 3.99 (2H, br s), 3.87 (3H, s), 3.66 (4H, br s), 2.41 (4H, br s), 2.32 (3H, s). 13C-NMR (100 MHz, CDCl3) δ:
171.1, 146.8, 138.5, 124.4, 121.0, 113.4, 110.3, 55.6, 54.6, 45.4. HRMS (ESI-TOF) m/z: [M+H]+ Calcd for
C13H20N3O2 250.1555; Found 250.1560.
5.1.12. 2-methoxy-4-(4-methylpiperazin-1-yl)aniline (20c)
A solution of commercially available 4-fluoro-2-methoxy-1-nitrobenzene (22) (200 mg, 1.17 mmol),
N-methyl piperazine (154 µL, 1.40 mmol) and Et3N (329 µL, 2.34 mmol) in DMA (2.00 mL) was stirred at
150 °C for 15 min utilizing a microwave reactor. The solution was purified by column chromatography
utilizing NH silica gel (DCM/Methanol) to afford 1-(3-methoxy-4-nitrophenyl)-4-methylpiperazine (282 mg,
96%).
To a solution of 1-(3-methoxy-4-nitrophenyl)-4-methylpiperazine (200 mg, 0.796 mmol) in methanol (5.00
mL) was added 10% Pd/C (50.0 mg) under nitrogen atmosphere and then filled with hydrogen (3 times). The
suspension was stirred at room temperature for 2 h. After the starting material consumption had been confirmed
by LC-MS, the reaction mixture was filtered with Celite® and concentrated. The obtained residue was purified
by column chromatography utilizing NH silica gel (DCM/methanol) to afford 20c (147 mg, 84%) as a dark
1
brown oil. H-NMR (400 MHz, DMSO-D6) δ: 6.51-6.48 (2H, m), 6.27 (1H, dd, J = 8.4, 2.6 Hz), 4.19 (2H, s),
3.73 (3H, s), 2.94-2.93 (4H, m), 2.43-2.42 (4H, m), 2.20 (3H, s). 13C-NMR (100 MHz, DMSO-D6) δ: 147.0,
143.3, 130.9, 114.2, 108.5, 101.8, 55.2, 54.9, 50.2, 45.8. HRMS (ESI-TOF) m/z: [M+H]+ Calcd for C12H20N3O
222.1606; Found 222.1618.
5.1.13. N-(2-methoxy-4-((3-(4-methylpiperazin-1-yl)propyl)carbamoyl)phenyl)-4-oxo-3,4-dihydrothieno
[3,2-d]pyrimidine-7-carboxamide (10)
To a solution of 4-oxo-3,4-dihydro-thieno[3,2-d]pyrimidine-7-carboxylic acid (19) (15.0 mg, 0.0760 mmol)
and 4-amino-3-methoxy-N-(3-(4-methylpiperazin-1-yl)propyl)benzamide (20b) (23.0 mg, 0.0760 mmol) in
acetonitrile (1.50 mL) was added HATU (29.0 mg, 0.152 mmol) and DIPEA (26.0 µL, 0.152 mmol) and stirred
at 80 °C. After the starting material consumption had been confirmed by LC-MS, the reaction mixture was
concentrated to give a crude residue. The obtained residue was purified by column chromatography utilizing

1
NH silica gel (DCM/methanol) to afford 10 (24.0 mg, 65%) as a light yellow solid. H-NMR (400 MHz,
DMSO-D6) δ: 12.94 (1H, s), 11.94 (1H, s), 8.92 (1H, s), 8.54-8.53 (2H, m), 8.45-8.44 (1H, m), 7.54 (1H, d, J =
1.5 Hz), 7.50 (1H, dd, J = 8.3, 2.0 Hz), 4.03 (3H, s), 3.28-3.26 (4H, br m), 2.42-2.35 (8H, br m), 2.15 (3H, s),
1.71-1.64 (2H, m). 13C-NMR (100 MHz, DMSO-D6) δ: 165.4, 158.7, 157.9, 152.9, 148.7, 148.2, 141.0, 130.3,
130.1, 129.8, 124.3, 119.8, 118.4, 109.7, 56.2, 55.7, 54.7, 52.7, 45.7, 37.9, 26.3. HRMS (ESI-TOF) m/z:
[M+H]+ Calcd for C23H29N6O4S 485.1971; Found 485.1995.
Compounds 11a-11l and 12 were synthesized by a similar procedure to 10.
5.1.14. N-(2-methyl-4-(4-methylpiperazine-1-carbonyl)phenyl)-4-oxo-3,4-dihydrothieno[3,2-d]pyrimidine
-7-carboxamide (11a)
The yield was 28%. White solid. 1H-NMR (600 MHz, TFA-D) δ: 9.33 (1H, s), 9.28 (1H, s), 7.55 (1H, d, J =
8.2 Hz), 7.48 (1H, s), 7.42 (1H, d, J = 8.0 Hz), 5.02 (1H, d, J = 14.3 Hz), 4.15 (1H, d, J = 14.6 Hz), 3.84-3.83
(2H, m), 3.70-3.68 (1H, m), 3.57-3.54 (1H, m), 3.36-3.34 (1H, m), 3.25-3.22 (1H, m), 3.10 (3H, s), 2.37 (3H, s).
13C-NMR (150 MHz, TFA-D) δ: 175.3, 163.4, 157.8, 151.3, 145.6, 145.5, 138.7, 137.9, 133.9, 132.2, 129.5,
127.9, 127.8, 125.7, 56.2, 47.0, 45.7, 42.1, 18.3. HRMS (ESI-TOF) m/z: [M+H]+ Calcd for C20H22N5O3S
412.1443; Found 412.1444.
5.1.15. N-(2-isopropyl-4-(4-methylpiperazine-1-carbonyl)phenyl)-4-oxo-3,4-dihydrothieno
[3,2-d]pyrimidine-7-carboxamide (11b)
1
The yield was 56%. White solid. H-NMR (400 MHz, DMSO-D6) δ: 12.99 (1H, br s), 11.47 (1H, s), 8.91
(1H, s), 8.48 (1H, s), 8.26 (1H, d, J = 8.2 Hz), 7.36 (1H, d, J = 2.0 Hz), 7.27 (1H, dd, J = 8.4, 2.0 Hz),
3.40-3.36 (5H, m), 2.34-2.31 (4H, m), 2.20 (3H, s), 1.31 (3H, s), 1.30 (3H, s). 13C-NMR (100 MHz, DMSO-D6)
δ: 168.9, 158.9, 158.2, 152.8, 148.9, 140.8, 138.3, 136.2, 131.6, 130.0, 125.1, 124.5, 124.3, 121.6, 54.5, 45.6,
27.5, 22.6. HRMS (ESI-TOF) m/z: [M+H]+ Calcd for C22H26N5O3S 4401756; Found 440.1758.
5.1.16. N-(5-(4-methylpiperazine-1-carbonyl)-[1,1'-biphenyl]-2-yl)-4-oxo-3,4-dihydrothieno
[3,2-d]pyrimidine-7-carboxamide (11c)
The yield was 97%. Light yellow solid. 1H-NMR (400 MHz, DMSO-D6) δ: 11.26 (1H, s), 8.86 (1H, s), 8.60
(1H, d, J = 8.3 Hz), 7.57-7.46 (7H, m), 7.26 (1H, d, J = 2.2 Hz), 3.52 (4H, br s), 2.32 (4H, br s), 2.19 (3H, s).
13C-NMR (100 MHz, DMSO-D6) δ: 168.4, 158.9, 158.7, 152.5, 148.1, 140.7, 137.4, 136.9, 132.5, 131.0, 129.7,
129.6, 129.0, 128.9, 127.9, 127.1, 123.9, 120.7, 54.6, 45.6. HRMS (ESI-TOF) m/z: [M-H]- Calcd for
C25H22N5O3S 472.1443; Found 472.1450.
5.1.17. N-(2-cyclohexyl-4-(4-methylpiperazine-1-carbonyl)phenyl)-4-oxo-3,4-dihydrothieno
[3,2-d]pyrimidine-7-carboxamide (11d)
1
The yield was 46%. White solid. H-NMR (400 MHz, DMSO-D6) δ: 13.02 (1H, s), 11.21 (1H, s), 8.97 (1H,

s), 8.42 (1H, s), 8.20 (1H, d, J = 8.4 Hz), 7.34 (1H, d, J = 1.8 Hz), 7.27 (1H, dd, J = 8.4, 2.0 Hz), 3.58-3.41 (4H,
m), 2.33 (4H, br s), 2.20 (3H, s), 1.86-1.83 (5H, m), 1.75-1.72 (1H, m), 1.47-1.44 (4H, m), 1.29-1.23 (1H, m).
13C-NMR (100 MHz, DMSO-D6) δ: 168.9, 158.7, 157.3, 152.7, 148.0, 141.4, 137.5, 135.9, 131.8, 130.1, 125.0,
124.5, 122.1, 54.5, 45.6, 37.7, 32.8, 26.5, 25.6. HRMS (ESI-TOF) m/z: [M+H]+ Calcd for C25H30N5O3S
480.2069; Found 480.2068.
5.1.18. N-(2-fluoro-4-(4-methylpiperazine-1-carbonyl)phenyl)-4-oxo-3,4-dihydrothieno
[3,2-d]pyrimidine-7-carboxamide (11e)
1
The yield was 78%. White solid. H-NMR (600 MHz, TFA-D) δ: 9.34 (1H, s), 9.29 (1H, s), 8.01 (1H, dd, J
= 7.8, 7.8 Hz), 7.42-7.41 (2H, m), 5.02 (1H, d, J = 13.4 Hz), 4.18-4.15 (1H, m), 3.86-3.85 (2H, m), 3.74-3.73
(1H, m), 3.58-3.55 (1H, m), 3.36-3.28 (2H, m), 3.12 (3H, s). 13C-NMR (150 MHz, TFA-D) δ: 173.7, 163.2,
158.1, 157.9, 156.4, 151.2, 145.8, 145.7, 133.7, 128.4, 128.0, 127.7, 126.0, 125.9, 56.1, 47.0, 45.7, 42.1.
HRMS (ESI-TOF) m/z: [M+H]+ Calcd for C19H19FN5O3S 416.1193; Found 416.1187.
5.1.19. N-(2-chloro-4-(4-methylpiperazine-1-carbonyl)phenyl)-4-oxo-3,4-dihydrothieno
[3,2-d]pyrimidine-7-carboxamide (11f)
The yield was 75%. White solid. 1H-NMR (600 MHz, TFA-D) δ: 9.36 (1H, s), 9.26 (1H, s), 8.03 (1H, br s),
7.71 (1H, s), 7.53 (1H, br s), 5.02-5.00 (1H, m), 4.14-4.12 (1H, m), 3.93-3.86 (2H, m), 3.75-3.73 (1H, m),
3.62-3.61 (1H, m), 3.34-3.26 (2H, m), 3.10 (3H, s). 13C-NMR (150 MHz, TFA-D) δ: 173.7, 163.2, 157.9, 151.4,
145.9, 145.4, 137.1, 133.9, 131.5, 131.3, 128.8, 128.4, 127.8, 126.1, 55.9, 47.1, 45.7, 42.2. HRMS (ESI-TOF)
m/z: [M+H]+ Calcd for C19H19ClN5O3S 432.0897; Found 432.0897.
5.1.20. N-(2-bromo-4-(4-methylpiperazine-1-carbonyl)phenyl)-4-oxo-3,4-dihydrothieno
[3,2-d]pyrimidine-7-carboxamide (11g)
The yield was 40%. White solid. 1H-NMR (600 MHz, TFA-D) δ: 9.34 (1H, s), 9.24 (1H, s), 7.99 (1H, d, J =
8.3 Hz), 7.86 (1H, s), 7.56 (1H, d, J = 8.4 Hz), 5.02-4.99 (1H, m), 4.15-4.13 (1H, m), 3.88-3.85 (2H, m),
3.73-3.71 (1H, m), 3.57-3.55 (1H, m), 3.36-3.27 (2H, m), 3.11 (3H, s). 13C-NMR (150 MHz, TFA-D) δ: 173.5,
163.2, 157.9, 151.3, 145.8, 145.3, 138.5, 134.4, 134.1, 129.4, 128.5, 127.8, 126.1, 121.0, 56.0, 47.0, 45.7, 42.2.
HRMS (ESI-TOF) m/z: [M +H]+ Calcd for C19H19BrN5O3S 476.0392; Found 476.0396.
5.1.21. N-(2-methoxy-4-(4-methylpiperazine-1-carbonyl)phenyl)-4-oxo-3,4-dihydrothieno
[3,2-d]pyrimidine-7-carboxamide (11h)
1
The yield was 97%. White solid. H-NMR (400 MHz, DMSO-D6) δ: 12.53 (1H, br s), 11.56 (1H, s), 8.85 (1H,
s), 8.49 (1H, d, J = 8.1 Hz), 8.37 (1H, s), 7.14 (1H, s), 7.08 (1H, d, J = 8.6 Hz), 4.00 (3H, s), 3.71 (4H, br s),
3.18 (4H, br s), 2.79 (3H, s). 13C-NMR (100 MHz, DMSO-D6) δ: 169.3, 159.1, 157.5, 153.3, 148.8, 148.4,
142.0, 130.6, 130.2, 129.9, 124.9, 120.4, 119.3, 110.6, 56.8, 52.9, 46.0, 43.0. HRMS (ESI-TOF) m/z: [M+H]+

Calcd for C20H22N5O4S 428.1393; Found 428.1400.
5.1.22. N-(2-isopropoxy-4-(4-methylpiperazine-1-carbonyl)phenyl)-4-oxo-3,4-dihydrothieno
[3,2-d]pyrimidine-7-carboxamide (11i)
1
The yield was 93%. White solid. H-NMR (400 MHz, DMSO-D6) δ: 12.99 (1H, br s), 11.68 (1H, s), 8.94
(1H, s), 8.55 (1H, d, J = 8.4 Hz), 8.33 (1H, s), 7.11 (1H, d, J = 1.6 Hz), 6.99 (1H, dd, J = 8.4, 1.6 Hz),
4.84-4.74 (1H, m), 3.50 (4H, br s), 2.32 (4H, br s), 2.20 (3H, s), 1.40 (3H, s), 1.39 (3H, s). 13C-NMR (100 MHz,
DMSO-D6) δ: 168.6, 158.5, 157.3, 152.7, 147.6, 146.1, 141.4, 131.2, 130.3, 129.8, 124.5, 119.4, 119.3, 112.1,
71.0, 54.5, 45.6, 21.9. HRMS (ESI-TOF) m/z: [M+H]+ Calcd for C22H26N5O4S 456.1705; Found 456.1709.
5.1.23. N-(2-((1,1,1,3,3,3-hexafluoropropan-2-yl)oxy)-4-(4-methylpiperazine-1-carbonyl)phenyl)-4-
oxo-3,4-dihydrothieno[3,2-d]pyrimidine-7-carboxamide (11j)
1
The yield was 21%. White solid. H-NMR (DMSO-D6) δ: 12.91 (1H, br s), 11.69 (1H, s), 9.00 (1H, s), 8.66
(1H, d, J = 8.3 Hz), 8.04 (1H, s), 7.62 (1H, s), 7.31 (1H, dd, J = 8.4, 1.6 Hz), 6.86-6.80 (1H, m), 3.63 (4H, br s),
3.03 (4H, br s), 2.69 (3H, s). 13C-NMR (100 MHz, DMSO-D6) δ: 167.9, 158.8, 156.8, 152.5, 146.9, 144.8,
142.2, 130.2, 129.6, 129.5, 128.8, 124.6, 123.2, 122.6, 120.6, 119.8, 112.8, 52.8, 43.0. HRMS (ESI-TOF) m/z:
[M+H]+ calcd for C22H20F6N5O4S 564.1140; Found 564.1141.
5.1.24. N-(2-(dimethylamino)-4-(4-methylpiperazine-1-carbonyl)phenyl)-4-oxo-3,4-dihydrothieno
[3,2-d]pyrimidine-7-carboxamide (11k)
The yield was 92%. White solid. 1H-NMR (DMSO-D6) δ: 12.01 (1H, s), 8.89 (1H, d, J = 1.6 Hz), 8.54 (1H,
d, J = 8.4 Hz), 8.50 (1H, s), 7.23 (1H, d, J = 1.8 Hz), 7.13 (1H, dd, J = 8.3, 1.9 Hz), 3.51 (4H, br s), 2.73 (6H,
s), 2.33 (4H, br s), 2.20 (3H, s). 13C-NMR (100 MHz, DMSO-D6) δ: 168.7, 158.6, 157.2, 152.9, 147.8, 143.5,
141.4, 134.4, 130.9, 130.5, 124.3, 122.8, 119.1, 54.5, 45.6, 44.0. HRMS (ESI-TOF) m/z: [M+H]+ Calcd for
C21H25N6O3S 441.1709; Found 441.1726.
5.1.25. N-(2-(ethyl(methyl)amino)-4-(4-methylpiperazine-1-carbonyl)phenyl)-4-oxo-3,4-dihydrothieno
[3,2-d]pyrimidine-7-carboxamide (11l)
1
The yield was 47%. White solid. H-NMR (400 MHz, DMSO-D6) δ: 12.47 (1H, br s), 11.72 (1H, s), 8.85
(1H, s), 8.53 (1H, d, J = 8.6 Hz), 8.29 (1H, s), 7.30 (1H, s), 7.21 (1H, d, J = 8.1 Hz), 3.72 (4H, br s), 3.07-2.96
(5H, m), 2.81 (3H, s), 2.70 (3H, s), 1.06 (3H, t, J = 7.0 Hz). 13C-NMR (100 MHz, DMSO-D6) δ: 168.9, 158.7,
157.1, 152.8, 147.5, 142.5, 141.7, 136.1, 130.5, 129.5, 124.4, 123.6, 121.2, 119.2, 52.3, 49.4, 42.4, 41.6, 12.3.
HRMS (ESI-TOF) m/z: [M+H]+ Calcd for C22H27N6O3S 455.1865; Found 455.1887.
5.1.26. N-(2-isopropoxy-3-(4-methylpiperazine-1-carbonyl)phenyl)-4-oxo-3,4-dihydrothieno
[3,2-d]pyrimidine-7-carboxamide (12)

The yield was 59%. White solid. 1H-NMR (400 MHz, CDCl3) δ: 11.39 (1H, s), 8.84 (1H, s), 8.65 (1H, dd, J
= 8.3, 1.5 Hz), 8.38 (1H, s), 7.19 (1H, dd, J = 8.1, 7.3 Hz), 7.07 (1H, dd, J = 7.3, 1.5 Hz), 4.45-4.39 (1H, m),
4.10-4.06 (1H, m), 3.72-3.69 (1H, m), 3.50-3.32 (2H, m), 2.62-2.59 (1H, m), 2.50-2.44 (2H, m), 2.35 (3H, s),
2.30-2.28 (1H, m). 13C-NMR (100 MHz, DMSO-D6) δ: 166.4, 158.6, 157.1, 152.5, 147.8, 142.5, 141.8, 132.6,
130.2, 129.9, 124.5, 123.9, 123.1, 121.2, 76.9, 54.6, 54.1, 46.3, 45.6, 41.0, 22.5, 21.3. HRMS (ESI-TOF) m/z:
[M+H]+ Calcd for C22H26N5O4S 456.1705; Found 456.1712.
5.2. In vitro TAK1 enzyme assay16
To screen TAK1 inhibitors, insect expression vectors against GST-TAK1 and His-TAB1 were infected into
insect cells. Then each tag-labeled kinase was purified using an anti-tag antibody-conjugated affinity column.
Each purified kinase was mixed and used as an enzyme for screening. The kinase enzyme was incubated with
various compounds and subsequently incubated with 6.25 µg of myelin basic protein (Upstate Cat.13-110) and
0.5 µCi of [33P]ATP (3,000 Ci/mmol, GE Healthcare) in 50 µL of the kinase buffer containing 25 mM TBS (pH
7.6), 0.5 mM dithiothreitol, 5 mM MgCl2, 0.0025% Tween20 at 30 °C for 30 min. Then 10 µL of 25%
trichloroacetic acid was added into the reaction well and samples were further incubated at 4 °C for 30 min.
Acid-precipitated samples were separated by a glass filter and liquid scintillation cocktails were added. The
radioactivity of 33P incorporated into myelin basic protein was measured with a MicroBeta (Perkinelmer).
5.3. Liver microsomal stability assay
An aliquot of 1 µM of each compound was incubated with human (or mouse) liver microsome (0.5 mg
protein/mL) in 50 mM phosphate buffer (pH 7.4) containing 1 mM NADPH (the reduced form of nicotinamide
adenine dinucleotide phosphate) at 37ꢀ°C for 30 min. After the enzyme reaction was quenched with the addition
of a three-fold volume of acetonitrile, the reaction mixture was centrifuged at 1500 rpm for 10 min. The
obtained supernatant was used as a test sample to measure the stability in human (or mouse) liver microsome
by measuring the compound in the sample using LC-MS/MS.
5.4. Solubility assay
An aliquot of 50 µL of 1 mM sample in DMSO solution was freeze-dried to remove DMSO. To the resulting
residue was added 50 µL of 50 mM phosphate buffer (pH 6.5) or FaSSIF solution, which was then irradiated
ultrasonically for 10 min. The solution was shaken for 24 h at 37 °C, centrifuged for 10 min at 25 °C (3000
rpm), and filtered by Whatman Unifilter. Concentration of the filtrate was analyzed by HPLC-UV based on the
calibration curve of each sample.
FaSSIF solution was prepared by the following procedure. A solution of L-α-phosphatidylcholine (0.295 g)

and sodium taurocholate (0.805 g) in purified water (5.00 mL) was stirred until the solution was clear. The
solution was diluted with purified water (450 mL) and added to KH2PO4 (1.95 g) and KCl (3.85g). The pH
value of this solution was adjusted to 6.5 by addition of 1N NaOH aq. and diluted with purified water until the
volume was 500 mL.
5.5. Expression and purification of TAK1-TAB1 fusion protein for crystallographic structure analysis
The recombinant fusion protein used for crystallization in this study consisted of human TAK1 kinase
domain (residue 31-303) and TAB1 (residue 468-504) and was prepared as described previously29. Briefly, the
truncated version of TAK1 fused to the TAB1 segment incorporated an N-terminal His6 tag and a HRV3c
protease cleavage site, and this was inserted into the bacterial/baculoviral expression vector pFastBac1
(Thermo Fisher Scientific). The enzyme was over-expressed in the baculovirus expression system using Sf9
insect cells and then purified from the cell extract using Ni affinity resin (GE Healthcare). Then the N-terminal
His6 tag was removed by PreScission protease (GE Healthcare) and the target enzyme was isolated by
size-exclusion chromatography on a Superdex 200 26/60 column (GE Healthcare). The resulting sample was
loaded onto an anion exchange column (HiTrap-Q HP 5mL, GE Healthcare) that had been equilibrated with 25
mM Tris, pH 8.0, 5 mM DTT, and 10%(v/v) glycerol and was then eluted using a linear gradient of 50–300
mM NaCl. Fractions of the forward peak corresponding to the non-phosphorylated form, as confirmed by
SDS-PAGE using Pro-Q Diamond Phosphoprotein Gel Stain (Thermo Fisher Scientific), were pooled and
−80
stored at
°C for crystallization.
5.6. Crystallization, data collection and structure determination
The purified TAK1-TAB1 fusion protein samples were concentrated to 3 mg/mL with about 0.2 mM of each
compound (10, 11c and 12), and crystallization was carried out by the hanging-drop vapor diffusion method at
20–21 °C. The reservoir solution consisted of 1.7 M sodium potassium phosphate (pH 7.7). Crystals appeared
and grew to full size within 4-6 days. These crystals were soaked in a solution of 1.7 M sodium potassium
phosphate (pH 7.7) containing 20% glycerol as a cryoprotectant. X-ray diffraction data from crystals of the
complex formed from TAK1-TAB1 with ligands 10, 11c, or 12 were collected at cryogenic temperatures by the
synchrotron radiation sources and detectors shown in Supplementary Table 1.
The diffraction datasets were indexed, integrated, and scaled using the program HKL200030 for the complex
of TAK1-TAB1 with ligand 10 or 11c, and using the autoPROC program31, XDS32, and Scala33 for the complex
of TAK1-TAB1 with ligand 12. The structure was determined by molecular replacement with the program
MolRep34 using the published TAK1-TAB1 crystal structure (PDB: 2EVA)29, as a search model. Models were
built with the program Coot35 and refined with the program Refmac536. The X-ray data collection and
refinement statistics are shown in Supplementary Table 1. Refined X-ray coordinates and structure factors of

TAK1-TAB1 in complex with ligands 10, 11c, and 12 have been deposited in the PDB with accession codes
5JGB, 5JGA, and 5JGD, respectively.
5.7. Modeling
To calculate the volume of R2 group of thienopyrimidinone analogs, the initial conformations of R2 groups
were manually modeled by changing R2 group of 11c (benzene ring) to each R2 group. The conformations were
optimized in the crystal structure of TAK1/11c complex (PDB: 5JGA) by the program Discovery Studio, v4.1
BioVia, San Diego, CA, USA with the TAK1 3D structure fixed. Volume of R2 group was calculated by using
the protocol “Calculate Molecular Properties” of Discovery Studio with the default settings.
Acknowledgments
We thank Y. Itezono for kind advice on NMR measurements, Y. Furuta and M. Arai for HRMS
measurements, and Y. Tachibana for kinase panel assay.
Reference and notes
1.
Yamaguchi, K.; Shirakabe, K.; Shibuya, H.; Irie, K.; Oishi, I.; Ueno, N.; Taniguchi,
T.; Nishida, E.; Matsumoto, K. Science (New York, N.Y.) 1995, 270, 2008.
2.
3.
Sakurai, H. Trends in pharmacological sciences 2012, 33, 522.
Ninomiya-Tsuji, J.; Kishimoto, K.; Hiyama, A.; Inoue, J.; Cao, Z.; Matsumoto, K.
Nature 1999, 398, 252.
4.
Sato, S.; Sanjo, H.; Takeda, K.; Ninomiya-Tsuji, J.; Yamamoto, M.; Kawai, T.;
Matsumoto, K.; Takeuchi, O.; Akira, S. Nature immunology 2005, 6, 1087.
5.
K.; Nishida, E.; Matsumoto, K. Science (New York, N.Y.) 1996, 272, 1179.
6. Takaesu, G.; Kishida, S.; Hiyama, A.; Yamaguchi, K.; Shibuya, H.; Irie, K.;
Ninomiya-Tsuji, J.; Matsumoto, K. Molecular cell 2000, 5, 649.
7. Ishitani, T.; Takaesu, G.; Ninomiya-Tsuji, J.; Shibuya, H.; Gaynor, R. B.;
Matsumoto, K. The EMBO journal 2003, 22, 6277.
8. Adcock, I. M.; Chung, K. F.; Caramori, G.; Ito, K. European journal of
pharmacology 2006, 533, 118.
Shibuya, H.; Yamaguchi, K.; Shirakabe, K.; Tonegawa, A.; Gotoh, Y.; Ueno, N.; Irie,
9.
Courties, G.; Seiffart, V.; Presumey, J.; Escriou, V.; Scherman, D.; Zwerina, J.; Ruiz,
G.; Zietara, N.; Jablonska, J.; Weiss, S.; Hoffmann, A.; Jorgensen, C.; Apparailly, F.; Gross, G.
Blood 2010, 116, 3505.
10.
Kumar, R.; Crouthamel, M. C.; Rominger, D. H.; Gontarek, R. R.; Tummino, P. J.;
Levin, R. A.; King, A. G. British journal of cancer 2009, 101, 1717.

11.
Izzedine, H.; Rixe, O.; Billemont, B.; Baumelou, A.; Deray, G. American journal of
kidney diseases : the official journal of the National Kidney Foundation 2007, 50, 203.
12.
13.
Gilliland, D. G.; Griffin, J. D. Blood 2002, 100, 1532.
Manning, G.; Whyte, D. B.; Martinez, R.; Hunter, T.; Sudarsanam, S. Science (New
York, N.Y.) 2002, 298, 1912.
14.
15.
Kilty, I.; Jones, L. H. Future medicinal chemistry 2015, 7, 23.
Wu, J.; Powell, F.; Larsen, N. A.; Lai, Z.; Byth, K. F.; Read, J.; Gu, R. F.; Roth, M.;
Toader, D.; Saeh, J. C.; Chen, H. ACS chemical biology 2013, 8, 643.
16. Ninomiya-Tsuji, J.; Kajino, T.; Ono, K.; Ohtomo, T.; Matsumoto, M.; Shiina, M.;
Mihara, M.; Tsuchiya, M.; Matsumoto, K. The Journal of biological chemistry 2003, 278,
18485.
17.
Bioscience, biotechnology, and biochemistry 2008, 72, 1894.
18. Kamiyama, H.; Kakeya, H.; Usui, T.; Nishikawa, K.; Shoji, M.; Hayashi, Y.; Osada,
H. Oncology research 2008, 17, 11.
Kamiyama, H.; Usui, T.; Sakurai, H.; Shoji, M.; Hayashi, Y.; Kakeya, H.; Osada, H.
19.
20.
Byth, K.; Palakurthi, S.; Yu, L.; Zhang, Q.; 2008.
Lockman, J. W.; Reeder, M. D.; Robinson, R.; Ormonde, P. A.; Cimbora, D. M.;
Williams, B. L.; Willardsen, J. A. Bioorganic & medicinal chemistry letters 2011, 21, 1724.
21. Buglio, D.; Palakurthi, S.; Byth, K.; Vega, F.; Toader, D.; Saeh, J.; Neelapu, S. S.;
Younes, A. Blood 2012, 120, 347.
22.
Hornberger, K. R.; Berger, D. M.; Crew, A. P.; Dong, H.; Kleinberg, A.; Li, A. H.;
Medeiros, M. R.; Mulvihill, M. J.; Siu, K.; Tarrant, J.; Wang, J.; Weng, F.; Wilde, V. L.;
Albertella, M.; Bittner, M.; Cooke, A.; Gray, M. J.; Maresca, P.; May, E.; Meyn, P.; Peick, W.,
Jr.; Romashko, D.; Tanowitz, M.; Tokar, B. Bioorganic & medicinal chemistry letters 2013,
23, 4517.
23.
Hornberger, K. R.; Chen, X.; Crew, A. P.; Kleinberg, A.; Ma, L.; Mulvihill, M. J.;
Wang, J.; Wilde, V. L.; Albertella, M.; Bittner, M.; Cooke, A.; Kadhim, S.; Kahler, J.; Maresca,
P.; May, E.; Meyn, P.; Romashko, D.; Tokar, B.; Turton, R. Bioorganic & medicinal chemistry
letters 2013, 23, 4511.
24.
Kilty, I.; Green, M. P.; Bell, A. S.; Brown, D. G.; Dodd, P. G.; Hewson, C.; Hughes, S.
J.; Phillips, C.; Ryckmans, T.; Smith, R. T.; van Hoorn, W. P.; Cohen, P.; Jones, L. H.
Chemical biology & drug design 2013, 82, 500.
25.
Arora, N.; Chen, S.; Hermann, J. C.; Kuglstatter, A.; Labadie, S. S.; Lin, C. J. J.;
Lucas, M. C.; Moore, A. G.; Papp, E.; Talamas, F. X.; Wanner, J.; Zhai, Y.; 2012.
26.
27.
Ishikawa, M.; Hashimoto, Y. Journal of medicinal chemistry 2011, 54, 1539.
Nishio, M.; Umezawa, Y.; Fantini, J.; Weiss, M. S.; Chakrabarti, P. Physical

chemistry chemical physics : PCCP 2014, 16, 12648.
28. Griffith, J.; Black, J.; Faerman, C.; Swenson, L.; Wynn, M.; Lu, F.; Lippke, J.;
Saxena, K. Molecular cell 2004, 13, 169.
29. Brown, K.; Vial, S. C.; Dedi, N.; Long, J. M.; Dunster, N. J.; Cheetham, G. M.
Journal of molecular biology 2005, 354, 1013.
30. Otwinowski, Z.; Minor, W. In Methods in Enzymology; Academic Press, 1997; Vol.
Volume 276, pp. 307.
31.
Vonrhein, C.; Flensburg, C.; Keller, P.; Sharff, A.; Smart, O.; Paciorek, W.; Womack,
T.; Bricogne, G. Acta crystallographica. Section D, Biological crystallography 2011, 67, 293.
32.
125.
33.
72.
34.
Kabsch, W. Acta crystallographica. Section D, Biological crystallography 2010, 66,
Evans, P. Acta crystallographica. Section D, Biological crystallography 2006, 62,
Vagin, A.; Teplyakov, A. Acta crystallographica. Section D, Biological
crystallography 2010, 66, 22.
35.
Emsley, P.; Lohkamp, B.; Scott, W. G.; Cowtan, K. Acta crystallographica. Section D,
Biological crystallography 2010, 66, 486.
36.
Gruene, T. Acta crystallographica. Section D, Biological crystallography 2013, 69,
1861.

*Graphical Abstract (for review)
structure based drug design
O
N
O
N
N
O
O
N
N
N
S
N
H
N
S
H
S
O
O
H
O
N
N
O
N
O
N
N
N
Metabolic identification
Hit compound
TAK1 inhibitory activity
37% inhibition@1μM
TAK1 IC50 = 2.3 nM
Low kinase selectivity
TAK1 IC50 = 11 nM
high kinase selectivity