Discovery of azabenzimidazole derivatives as potent, selective inhibitors of TBK1/IKK kinases

Article abstract of DOI:10.1016/j.bmcl.2012.01.018

The design, synthesis and biological evaluation of a series of azabenzimidazole derivatives as TBK1/IKK kinase inhibitors are described. Starting from a lead compound 1a, iterative design and SAR exploitation of the scaffold led to analogues with nM enzyme potencies against TBK1/IKK. These compounds also exhibited excellent cellular activity against TBK1. Further structure-based design to improve selectivity over CDK2 and Aurora B resulted in compounds such as 5b-e. These probe compounds will facilitate study of the complex cancer biology of TBK1 and IKK.

Full text of DOI:10.1016/j.bmcl.2012.01.018

Bioorganic & Medicinal Chemistry Letters 22 (2012) 2063–2069
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of azabenzimidazole derivatives as potent, selective inhibitors
of TBK1/IKKe kinases
Tao Wang a, , Michael A. Block a, Scott Cowen a, Audrey M. Davies a, Erik Devereaux b,
⇑
Lakshmaiah Gingipalli a, Jeffrey Johannes a, Nicholas A. Larsen b, Qibin Su a, Julie A. Tucker c,
David Whitston b, Jiaquan Wu b, Hai-Jun Zhang a, Michael Zinda b, Claudio Chuaqui a,
⇑
a Department of Cancer Chemistry, Oncology Innovative Medicines Unit, AstraZeneca R&D Boston, 35 Gatehouse Drive, Waltham, MA 02451, USA
b Department of Cancer Bioscience, Oncology Innovative Medicines Unit, AstraZeneca R&D Boston, 35 Gatehouse Drive, Waltham, MA 02451, USA
c Discovery Sciences, AstraZeneca R&D, Alderley Park, Cheshire, United Kingdom
a r t i c l e i n f o
a b s t r a c t
Article history:
The design, synthesis and biological evaluation of a series of azabenzimidazole derivatives as TBK1/IKK
kinase inhibitors are described. Starting from a lead compound 1a, iterative design and SAR exploitation
e
Received 11 November 2011
Revised 3 January 2012
Accepted 9 January 2012
Available online 14 January 2012
of the scaffold led to analogues with nM enzyme potencies against TBK1/IKK . These compounds also
e
exhibited excellent cellular activity against TBK1. Further structure-based design to improve selectivity
over CDK2 and Aurora B resulted in compounds such as 5b–e. These probe compounds will facilitate
study of the complex cancer biology of TBK1 and IKKe.
Keywords:
Ó 2012 Elsevier Ltd. All rights reserved.
TBK1
IKKe
Azabenzimidazole
Kinase inhibitor
Protein kinases are important members of the encoded human
genome that are involved in the regulation of normal and patho-
physiological processes.1,2 One sub-family of the ca. 500 membered
HN
N
N
7
Br
N
protein kinases is the I
IKK , IKKb, IKK , IKK and TBK1. These I
ical components in the NF- B activation pathway that leads to
j
B kinases which consists of five isoforms:
2
6
O
a
c
e
jB kinases serve as the crit-
N
H
j
RELA/B activation and the regulation of cell proliferation, survival,
1a
innate and adaptive immune responses.3 Among these five mem-
Figure 1. Azabenzimidazole hit from screening.
bers, IKK
extensively studied and targeted using small molecule inhibitors
for the treatment of cancer and other diseases.4 Besides the NF-
pathway, TBK1 and IKK were shown to be essential components
of the interferon regulatory factor 3 (IRF3) signaling pathway.5
TBK1 and IKK have attracted great interest in the cancer area in
recent years. It was reported that TBK1 kinase activity was up-
regulated in certain transformed cells and that TBK1 was required
for the survival of these cells in culture.6,7 An integrative genomic
approach that incorporated an overexpression screen, RNAi screen
c lacks kinase domain while IKKa and IKKb have been
induced cell death.8 A recent report showed that IKK
e and TBK1
jB
e
act together by phosphorylating AKT in a PI3K signaling dependent
manner and that the dual phosphorylation resulted in a robust AKT
activation in vitro.9 So far there have been few reports on small
e
molecule inhibitors of TBK1 and/or IKKe 10–12
.
We herein report
the discovery of a novel series of azabenzimidazole derivatives as
potent and selective TBK1/IKK kinase inhibitors.
e
Screening of a kinase focused subset of the AstraZeneca com-
pound collection against the TBK1 enzyme identified compound
1a as an initial hit (Fig. 1). Compound 1a had an IC50 of 0.151
against the TBK1 enzyme. Compound 1a was also active against
and comparative genomic analysis identified IKK
e as an oncogene
in human breast cancer, and showed that suppression of IKK
e
lM
expression in breast cancer cell lines with IKK
e amplifications
IKKe (IC50: 0.707 lM). A homology model of TBK1 was built using
Prime v.1.513 based on the structure of PAK4 (AstraZeneca proprie-
tary structure, unpublished data). A proposed binding mode for the
azabenzimidazole scaffold was generated using Glide v.4.514 where
the inhibitor binds to the hinge of the kinase via the pyridine
⇑
Corresponding authors. Tel.: +1 781 839 4734; fax: +1 781 839 4230 (T.W.); tel.:
+1 781 839 4106; fax: +1 781 839 4230 (C.C.).
E-mail addresses: tao.wang@astrazeneca.com (T. Wang), claudio.chuaqui@
astrazeneca.com (C. Chuaqui).
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2012.01.018

2064
T. Wang et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2063–2069
Table 1
Initial 6- and 7-position SAR
R7
N
R6
N
O
N
H
N
1a-h
Compd
R6
R7
TBK1 IC50
0.151
(
lM)
IKKe IC50 (lM)
Me
1a
Br
Br
0.707
NT
HN
1b
1c
1d
1.22
2.29
10.3
N
H
N
Br
Br
NT
NT
N
H
N
N
H
HN
1e
1f
Br
Br
0.105
0.468
0.523
4.32
N
H
HN
Me
NH
1g
1h
CN
Br
0.426
2.74
NT
NT
HN
Cl
nitrogen and the imidazole N–H groups while the piperidine side
chain is directed to the solvent. The 6-substituent (Br on 1a) points
(1c and 1d). Interestingly, piperidine chains of variable length were
tolerated in TBK1 (1e and 1f). Compound 1e also has comparable
potency against IKK
is a matched pair of 1a and showed an IC50 of 0.426 lM against
TBK1, suggesting that 6-cyano was at least as good as a 6-Br group.
When the amino group was replaced with a simple Cl, the resulting
analogue 1h was much less potent, demonstrating the importance
of the amino group at the 7-position. We conclude from this 7-
position exploration that a methylamino group was one of the best
groups in terms of TBK1 enzyme potency and it was kept constant
to allow further SAR work at other positions.
An SAR study of the 2-position of the azabenzimidazole was car-
ried out next. As illustrated in Table 2, removal of the methylenepi-
peridine moiety from 1a gave 2a with a comparable IC50 of
to the gatekeeper methionine (conserved in TBK1 and IKK
e
) (Fig. 2).
e
as compared to that of 1a.15 Compound 1g
This chemical series was attractive as three positions (2, 6 and
7-positions) on compound 1a offered clear opportunities to devel-
op structure–activity relationships. First, the substitution at the
2-position was fixed with a para-substituted phenyl moiety and
the R6 group was kept constant as Br (or CN as one matched pair)
to allow the probing of SAR off the amino group at the 7-position.
As the data in Table 1 show, replacement of the methyl group on 1a
with a more bulky lipophilic group (cyclopentyl on 1b) resulted in
an eightfold decrease in TBK1 enzyme activity.15 Linear chains of
variable chain length and with a polar terminal di-methylamino
group also led to a significant drop in TBK1 enzyme potency
0.335 lM against the TBK1 enzyme. Both electron-donating and
electron-withdrawing functional groups were tolerated at the 40-
position such as –OEt (2b), –OH (2d), –CO2H (2e) and –SO2Me
(2f), with the exception of –OiPr (2c). Moving the methyl sulfone
to the meta-position on the phenyl ring resulted in a drop of enzyme
potency to 9.55
lM. However, both meta- and para-NH2 analogues
had reasonable potencies of 0.065 (2h) and 0.111
lM (2i), respec-
tively. The phenyl ring itself could be replaced with an aminopyri-
dine without losing activity (2j). Bis-substitution (such as 2k)
retained the TBK1 potency. A variety of basic side chains besides
the original piperidine on 1a were also tolerated (2l–r), suggesting
that this area offers potential for modulating the physical proper-
ties of the inhibitor without affecting the target activities. Replace-
ment of the phenyl ring with another aromatic system yielded both
active (2s, indazole) and less active (2t, N-Me imidazole) analogues.
As expected from the high sequence/binding site similarities be-
tween TBK1 and IKK
also measured, both enzyme potencies (TBK1 and IKK
e
, for those analogues whose IKK
e
IC50’s were
) tracked
e
well with one another.
To determine whether the bromo was optimal at the 6-position
of the azabenzimidazole, additional substitutions were explored
beyond Br and cyano groups (Table 3). It was found that the
6-CONH2 moiety gave an analogue with an equal TBK1 enzyme po-
Figure 2. Azabenzimidazole hit 1a in a TBK1 homology model.

T. Wang et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2063–2069
2065
Table 2
SAR of the 2-position of benzimidazoles
HN
Br
N
R2
2a-t
R2
N
H
Compd
N
Compd
R2
TBK1 IC50
0.335
(
lM)
IKK
e
IC50
(
lM)
TBK1 IC50
0.345
(l
M)
IKK
e IC50 (lM)
OMe
OMe
OMe
2a
NT
2k
NT
2b
2c
0.231
1.41
NT
NT
2l
0.388
0.213
5.96
NT
OEt
O
N
N
O
2m
N
O
OH
2d
2e
0.201
0.257
NT
2n
2o
0.233
0.178
NT
NT
O
O
O
OH
O
0.590
NT
N
N
O
N
O
S
2f
0.229
9.55
2p
2q
0.106
0.196
NT
N
O
O
S
O
O
2g
>30
0.627
0.51
N
N
H
N
2h
2i
0.065
0.111
0.093
0.093
0.425
0.120
2r
2s
2t
N
0.192
0.288
1.52
N
NH2
NH2
N
NT
N
H
N
N
2j
2.46
N
NH2
Table 3
recapitulates the response of viral RNA) and then treated with
potential inhibitors to assess the inhibitory activity. When com-
bined with 2-(para-MeO)-phenyl and 6-Br, both amino and oxygen
linked piperidine side chains produced analogues with TBK1 IC50s
below 100 nM for the first time in this series (4a–b in Table 4).
SAR of the 6-position of benzimidazoles
O
HN
Their IC50s against IKK
TBK1 enzyme to cell drop off was within 2 to 3-fold.
e
were just ꢀ4 to 5-fold less active and the
N
R6
N
The PDK1 inhibitors BX-320 and BX-79512 have been reported to
have nM TBK1/IKKe activity.11 From binding mode overlays in TBK1
between the azabenzimidazole scaffold (represented by 3a) and
BX-320 (Fig. 3, orange) it was reasoned that similar side chains
should be evaluated at the 7-position of the azabenzimidazole scaf-
fold. Indeed, amido side chains were reported previously on TBK1/
O
N
N
3a-c
H
Compd
R6
TBK1 IC50
(
l
M)
IKKe IC50 (lM)
3a
2o
3b
3c
–CONH2
0.162
0.178
0.383
12
NT
NT
16.1
NT
inhibitors.12 An analogue with 30-acetyl-10,30-diamine at the 7-
M. Changing the terminal
group of the 7-position from Me to Pr boosted the TBK1 enzyme
potency to 0.010 M (4d) (Table 4). Docking studies suggest that
Br
Cl
H
IKKe
position (4c) had a TBK1 IC50 of 0.070
l
i
l
tency (3a). Replacement of the Br with Cl (3b) resulted in a slight
TBK1 enzyme potency decrease. Removal of the Br, however, ren-
dered a loss of TBK1 activity (3c).
After exploration of the 2- and 6-positions of the azabenzimi-
dazole, the SAR of the 7-position was revisited. At this point, a
TBK1 reporter assay was developed to assess the cellular potency
of these compounds.16 In this cell assay, the stably transfected
amide substituents interact with a lipophilic region of P-loop and
the alkyl side chain of the conserved Lys. The improved potency
of bulkier hydrophobic groups may be rationalized by increased
hydrophobic interactions between the inhibitor and the P-loop
sub-pocket. Potencies against TBK1 cell and IKK
e enzyme were also
improved to 0.014 and 0.084 M, respectively. A bulky alkyl termi-
l
t
nal group (e.g., Bu in 4e) was also tolerated. Regarding cycloalkyl
terminal groups, all 3–5-membered rings (4f–k) were potent and
the cyclobutyl analogue 4h had single digit nM potencies against
HEK293-TLR3 NF-
jB ISRE-Luciferase cell line was stimulated with
its putative ligand (in this case a synthetic dsRNA POLY I:C which

2066
T. Wang et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2063–2069
Table 4
SAR of the 7-position of azabenzimidazoles
R7
Br
N
R2
4a-p
N
H
N
Compd R7
R2
TBK1 IC50
0.057
(
l
M) IKK
e
IC50
(l
M) TBK1 cell EC50
(
l
M) Aurora B IC50
(l
M) CDK2 IC50
0.135
(lM)
NH
H
OMe
4a
0.265
0.172
0.111
0.043
N
NH
4b
0.039
0.070
0.076
0.073
0.031
0.024
0.133
0.370
O
H
N
H
N
4c
4d
4e
4f
0.543
0.084
0.060
0.115
0.005
0.021
0.020
O
O
O
O
H
N
H
N
0.010
0.007
0.018
0.004
0.004
0.005
0.014
0.011
0.025
0.020
0.006
0.006
0.030
0.081
0.025
0.029
0.038
0.007
>5.28
>30
H
N
H
N
H
N
H
N
1.37
H
N
H
N
NH2
4g
4h
4i
0.070
>4.91
0.187
O
O
O
H
N
H
N
H
N
N
OH
H
N
H
N
4j
4k
4l
0.008
0.010
0.106
0.046
0.076
0.040
0.008
0.084
0.016
0.093
0.733
0.345
2.79
3.97
O
O
H
N
H
N
H
N
H
N
14.9
O
O
N
H
N
H
N
4m
4n
0.059
0.046
0.278
0.541
0.087
0.026
NT
NT
NT
NT
S
H
N
H
N
O
O
O
H
N
H
N
NH2
4o
4p
0.003
0.004
0.004
0.017
0.004
0.012
0.004
0.013
0.155
0.518
H
N
H
N
O
N
both TBK1 enzyme and cell. The hydrogen atom on the amido nitro-
gen was not necessary as shown with analogue 4i (N-Me). Aryl
analogues, including benzoic (4l) and nicotinic (4m) were slightly
less potent as compared to the cyclobutyl analogues. When the
para-MeO phenyl was replaced with other preferred groups based
on the data shown in Table 2, highly potent analogues such as 4o
and 4p were obtained. Compound 4d was profiled against 79 ki-
nases (single point at 1 lM) in order to assess the selectivity of

T. Wang et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2063–2069
2067
Figure 3. Overlay of azabenzimidazole 3a with BX-320.
the azabenzimidazole series. The Gini coefficient computed from
these data is 0.556 which indicates the compound is moderately
selective with six kinases being inhibited within 10-fold of
TBK1.17 As structurally related compounds were reported as weak
Aurora inhibitors in the literature,18 this series was profiled against
Aurora B and CDK2 enzymes as counter screens and the data
showed that most of the 7-substituted analogues in Table 4 (except
Figure 4. Compound 4g in a CDK2 crystal structure.
As both CDK2 and Aurora B kinases are important players in cell
cycle progression and regulation, inhibition of those kinases would
complicate our study of the effects TBK1/IKKe modulation on can-
cer biology. It was therefore desirable to improve the selectivity
of our series over CDK2 and Aurora B kinases. A comparative bind-
ing site sequence and structural analysis between TBK1/IKK
and Aurora/CDK2 revealed several potential target sites on the
protein to drive selectivity over Aurora/CDK for this series. More-
for 4l , which was also less potent against TBK1/IKKe) were potent
inhibitors against Aurora B kinase at least at the enzyme evel. A
e
majority of these analogues also had IC50s of less than 1
CDK2.
lM against
Table 5
Improvement of kinase selectivity over Aurora B and CDK2
R7
R6
N
R2
5a-e
N
H
N
Compd R6
R7
R2
TBK1 IC50
M)
IKK
e
IC50
TBK1 cell EC50
M)
Aurora B IC50
M)
CDK2 IC50
(
l
(l
M)
(
l
(
l
(l
M)
H
N
H
N
5a
5b
CONH2
0.004
0.009
0.009
1.46
0.478
1.58
9.41
O
N
O
O
H
N
H
N
OMe
CONH2
Br
0.046
0.102
0.139
>30
OH
H
N
H
N
5c
0.032
0.035
0.872
4.65
MeO
O
O
H
N
H
N
F
5d
5e
Br
Br
0.04
0.122
0.204
0.017
0.034
0.37
6.52
2.02
MeO
MeO
H
N
H
N
F
0.038
0.664
O

2068
T. Wang et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2063–2069
O
O
OH
Cl
H2N
NH2
Cl
N
O
Br
NH2
NH2
9
7
HO
Br
N
a
b
N
N
H
6
8
O
NH2
HN
N
HO
HN
HO
HN
c
O
O
Br
Br
N
N
H
N
H
N
N
5c
10
Scheme 1. Reagents and conditions: (a) 2-methoxybenzoic acid (7), POCl3, CH3CN, microwave, 150 °C, 1 h; (b) 2-hydroxyl-1,3-diamine (9, excess), n-BuOH, microwave,
160 °C, 4 h; (c) cyclobutanecarboxylic acid, HATU, Et3N, DMF, room temperature, 2 h.
over, an X-ray crystal structure of 4g in complex with CDK2 aided
the design of inhibitors with reduced CDK2 potency (Fig. 4).19 For
the purpose of tuning the azabenzimidazole scaffold selectivity,
the most obvious residue to exploit is the gatekeeper on the hinge
Acknowledgments
The authors would like to thank Judith Stanway for insect cell
culture, Anna Valentine for protein purification, Claire Brassington
for protein crystallization, Scott Williams and Melissa Vasbinder
for the helpful discussions.
which is variable for the off-target kinases: TBK1 (Met), IKKe (Met),
Aurora family (Leu), and CDK2 (Phe). As shown in Figure 4, the
6-substituent of the azabenzimidazole points to the gatekeeper,
therefore it was reasoned that modification off the 6-position might
modulate the selectivity against the Aurora B and CDK2 kinases. As
the data in Table 5 show, analogues with a 6-CONH2 group (5a–b)
exhibited significant improvements in selectivity over Aurora B
References and notes
1. Dancey, J.; Sausville, E. A. Nat. Rev. Drug Disc. 2003, 2, 296.
2. Zhang, J.; Yang, P. L.; Gray, N. S. Nat. Rev. Cancer 2009, 9, 28.
3. Lee, D.-F.; Hung, M.-C. Clin. Cancer Res. 2008, 14, 5656.
4. Burke, J. R. Curr. Opin. Drug Disc. Dev. 2003, 6, 720.
5. Fitzgerald, K. A.; McWhirter, S. M.; Faia, K. L.; Rowe, D. C.; Latz, E.; Golenbock, D.
T.; Coyle, A. J.; Liao, S.-M.; Maniatis, T. Nat. Immunol. 2003, 4, 491.
6. Chien, Y.; Kim, S.; Bumeister, R.; Loo, Y.-M.; Kwon, S. W.; Johnson, C. L.;
Balakireva, M. G.; Romeo, Y.; Kopelovich, L.; Gale, M., Jr.; Yeaman, C.; Camonis,
J. H.; Zhao, Y.; White, M. A. Cell 2006, 127, 157.
and CDK2 while retaining TBK1 and IKK
The TBK1 cellular potency for 5a dropped to 1.46
due to poor permeability, while analogue 5b also gave sub-optimal
potency (0.139 M) in the TBK1 cell assay. Our second strategy to
e
enzymatic potencies.
l
M presumably
l
gain selectivity was through modification at the solvent channel
(2-substituent of the azabenzimidazole) since there are also differ-
ences among these kinases in this region (more prominent between
TBK1/IKKe and CDK2). After some library exploration work, it was
discovered that the analogue with an ortho-MeO phenyl at the 2-
7. Bodemann, B. O.; White, M. A. Nat. Rev. Cancer 2008, 8, 133.
8. Boehm, J. S.; Zhao, J. J.; Yao, J.; Kim, S. Y.; Firestein, R.; Dunn, I. F.; Sjostrom, S. K.;
Garraway, L. A.; Weremowicz, S.; Richardson, A. L.; Greulich, H.; Stewart, C. J.;
Mulvey, L. A.; Shen, R. R.; Ambrogio, L.; Hirozane-Kishikawa, T.; Hill, D. E.;
Vidal, M.; Meyerson, M.; Grenier, J. K.; Hinkle, G.; Root, D. E.; Roberts, T. M.;
Lander, E. S.; Polyak, K.; Hahn, W. C. Cell 2007, 129, 1065.
position and a 7-side chain containing a hydroxyl group (5c) had
9. Xie, X.; Zhang, D.; Zhao, B.; Lu, M.-K.; You, M.; Condorelli, G.; Wang, C.-Y.; Guan,
K.-L. Proc. Natl. Acad. Sci. U.S.A. 2011, 108, 6474.
a balanced potency profile: 0.032/0.102
l
M against TBK1/IKK
e;
0.035 M in TBK1 cell assay, 0.870/4.65
l
lM against Aurora B/
10. Bamborough, P.; Christopher, J. A.; Cutler, G. J.; Dickson, M. C.; Mellor, G. W.;
Morey, J. V.; Patel, C. B.; Shewchuk, L. M. Bioorg. Med. Chem. Lett. 2006, 16, 6236.
11. Bain, J.; Plater, L.; Elliott, M.; Shpiro, N.; Hastie, C. J.; Mclauchlan, H.; Klevernic,
I.; Simon, C.; Arthur, J. S. C.; Alessi, D. R.; Cohen, P. Biochem. J. 2007, 408, 297.
12. Feldman, R. I.; Wu, J. M.; Polokoff, M. A.; Kochanny, M. J.; Dinter, H.; Zhu, D.;
Biroc, S. L.; Alicke, B.; Bryant, J.; Yuan, S.; Buckman, B. O.; Lentz, D.; Ferrer, M.;
Whitlow, M.; Adler, M.; Finster, S.; Chang, Z.; Arnaiz, D. O. J. Biol. Chem. 2005,
280, 19867.
CDK2. Close analogues 5d–e also followed the same trend. As
shown above, our structure-based design approach took advantage
of the differences between TBK1/IKKe and CDK2/Aurora B at both
the gatekeeper position and the solvent channel and resulted in po-
tent and selective compounds 5b–e.
A representative synthesis of the lead compound 5c of this ser-
ies is illustrated in Scheme 1. The synthesis commenced with a tet-
ra-substituted pyridine 6. Treatment of 6 with 7 and POCl3 in
CH3CN gave 8 with the azabenzimidazole core structure installed.
Nucleophilic aromatic substitution at the 7-position of azabenzim-
idazole 8 with excess 2-hydroxyl-1,3-diamine 9 produced interme-
diate 10, which then underwent an amide coupling reaction to
yield the target compound 5c.
In summary, a series of novel azabenzimidazole analogues was
discovered and optimized against TBK1/IKKe kinases. Structure-
based design targeting the gatekeeper residues and the solvent
channel resulted in lead compounds (such as 5b–e) that showed
13. Prime, version 1.5; Schrodinger, LLC, New York, NY, 2010.
14. Glide, version 4.5, Schrodinger, LLC, New York, NY, 2010. The homology model
structure was prepared using the default protein preparation workflow. Grids
were generated using the centroid of an arbitrary ligand placed in the ATP
binding site incorporating hydrogen bonding constraints to the hinge. Ligands
were prepared using proprietary in-house software. Docking was carried out
using the XP option and constrained to make both hydrogen bonds to the hinge
of the kinases. Five poses were retained for each ligand.
15. TBK1/IKK
e
enzyme assays: Inhibitory effect of compounds against TBK1 and
typical experiment, series of kinase
IKK was evaluated by IC50s. In
e
a
a
catalyzed reactions with different concentrations of an inhibitor were set up in
HEPES buffer with a kinase and its substrates (for TBK1, 3 nM recombinant full-
length TBK1 (Life Technologies, Madison. WI), 1.8
KRRRAL(pS)VASLPGL, Primm Biotech, Cambridge, MA), 30
MgCl2; for IKK 4 nM IKK 1.5 peptide substrate (5FAM-
AKELDQGSLCTpSFVGTLQ-NH2, 21st Century Biochemicals, Marlborough,
MA), ATP, and 10 mM MgCl2). For 10-point IC50 test, inhibitor
concentrations were usually started from 10 M, followed by nine half-log
lM CK1tide (5FAM-Ahx-
l
M ATP, and 10 mM
e
,
e,
lM
good TBK1 cellular activity and were selective for TBK1/IKKe over
Aurora B/CDK2 kinases. These compounds could serve as useful
5 lM
l
dilutions. Reactions were kept at room temperature for a pre-defined duration
of time, quenched using HEPES buffer containing EDTA, and then analyzed
probe compounds in the study of the complex cancer biology of
TBK1 and IKKe kinases.

T. Wang et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2063–2069
2069
using microfluidic mobility shift assay on a Caliper LC3000 system. The ratio of
product (phosphorylated peptide) to total peptide was used to calculate
percentage inhibition for each inhibitor concentration. IC50s were obtained by
18. Bavetsias, V.; Sun, C.; Bouloc, N.; Reynisson, J.; Workman, P.; Linardopoulos, S.;
McDonald, E. Bioorg. Med. Chem. Lett. 2007, 17, 6567.
19. Crystallization, Structure Solution and Refinement: Recombinant protein and
crystals were obtained as described (Lawrie, A.M., et al., Nat. Struct. Biol. 1997,
4, 796–800). Crystals were cross-linked by incubation in mother liquor
containing 0.14% glutaraldehyde for 15 min before soaking in 20 mM 4g
overnight in mother liquor containing 20% DMSO. Diffraction data were
collected to 2.0 Å resolution in-house on a Rigaku MicroMax007 rotating anode
source equipped with a Saturn92 CCD detector. Data were integrated and
scaled using d⁄Trek software (Pflugrath, J. W. Acta Crystallogr., D 1999, 55,
1718–1725). The structure was determined by molecular replacement using
AMoRe (Bailey, S. Acta Crystallogr., D 1994, 50, 760–763) and refined using
Refmac with rebuilding in Coot (Emsley, P., et al. Acta Crystallogr., D, 2010, 66,
486–501). Waters were added to the structure using Buster (Bricogne G, et al.,
2009, BUSTER version 2.8.0. Cambridge United Kingdom: Global Phasing Ltd.)
and removed after visual inspection of the electron density. Final refinement
statistics converged with Rwork = 0.234 and Rfree = 0.303. The final density
surrounding the inhibitor was excellent. The coordinates and structure factors
have been deposited in the protein data bank under accession code 3ULI.
fitting the percentage inhibition data to
equation using Abase (IDBS).
a four-parameter sigmoidal IC50
16. TBK1 cell assay: On day 1, plate 40000 HEK293-TLR3 NF-
jB-ISRE-Luciferase
cells/well for a final volume 100 L in Corning flat bottomed 96 well poly-D-
l
lysine coated plates in DMEM containing 10% FBS. On day 2 prepare poly I:C,
(1 mg/mL stock in 1 ꢁ PBS) by heating at 50 °C, mix thoroughly, use DMEM
containing 0.5% FBS to dilute to a final concentration of 50 lg/mL poly I:C. The
synthetic dsRNA poly I:C recapitulates the response of viral RNA, stimulating
TLR3. Aspirate media from cell plates, replace with 90 L DMEM containing
0.5% FBS in column 2 and 90 L DMEM containing 0.5% FBS + 50 g/mg poly I:C
in all remaining wells. Compounds diluted to a final working concentration of
30 M–0.005 M on cell plates, dosed in triplicate, two columns remain for
stimulation with poly I:C, and no stimulation control. Incubate 4.5 h at 37 °C,
then add 100 L of SteadyGlo to each well in the dark. Incubate 15 min.
Transfer entire contents of each well, (200 L), to Corning flat bottomed 96
l
l
l
l
l
l
l
well white plates. Plates read on Tecan Ultra, data analyzed in Excel.
17. Graczyk, P. P. J. Med. Chem. 2007, 50, 5773.