Displacement assay for the detection of stabilizers of inactive kinase conformations

Article abstract of DOI:10.1021/jm901297e

Targeting protein kinases with small molecules outside the highly conserved ATP pocket to stabilize inactive kinase conformations is becoming a more desirable approach in kinase inhibitor research, since these molecules have advanced pharmacological prope

Full text of DOI:10.1021/jm901297e

J. Med. Chem. 2010, 53, 357–367 357
DOI: 10.1021/jm901297e
Displacement Assay for the Detection of Stabilizers of Inactive Kinase Conformations†
‡
Sabine Kluter, Christian Grutter, Tabassum Naqvi, Matthias Rabiller, Jeffrey R. Simard, Vijaykumar Pawar,
‡
§
‡
‡
‡
€
Matthaus Getlik, and Daniel Rauh*
€
‡
,‡
€
‡Chemical Genomics Centre of the Max Planck Society, Otto-Hahn-Strasse 15, D-44227 Dortmund, Germany and
§DiscoveRx Corporation, 42501 Albrae Street, Fremont, California 94538
Received August 30, 2009
Targeting protein kinases with small molecules outside the highly conserved ATP pocket to stabilize
inactive kinase conformations is becoming a more desirable approach in kinase inhibitor research, since
these molecules have advanced pharmacological properties compared to compounds exclusively
targeting the ATP pocket. Traditional screening approaches for kinase inhibitors are often based on
enzyme activity, but they may miss inhibitors that stabilize inactive kinase conformations by enriching
the active state of the kinase. Here we present the development of a kinase binding assay employing a
pyrazolourea type III inhibitor and enzyme fragment complementation (EFC) technology that is
suitable to screen stabilizers of enzymatically inactive kinases. To validate this assay system, we report
the binding characteristics of a series of kinase inhibitors to inactive p38R and JNK2. Additionally, we
present protein X-ray crystallography studies to examine the binding modes of potent quinoline-based
DFG-out binders in p38R.
Introduction
conformation. Type II inhibitors such as imatinib and sor-
afenib (25)9 also occupy this site but, in addition, extend
into the ATP binding region of the catalytic cleft to form
additional interactions with the hinge region of the kinase
domain.10 Type II/III inhibitors not onlyare perceivedtohave
superior selectivity profiles but also, because of their slow
off-rates and increased drug-target residence times, have
advanced pharmacological properties.10 Despite the advan-
tages of such ligands, methods for identifying type II/III
inhibitors are limited and most are unsuitable for high-
throughput screening (HTS) initiatives. Additionally, there
is no simple method available at all to predict whether a kinase
of interest can even adopt the DFG-out conformation.
Existing kinase screening methods are mainly based on
enzyme activity assays.11,12 Since active kinases are typically
in the DFG-in conformation, these assays create a bias for the
identification of classic type I inhibitors and, therefore, may not
be particularly suitable for detecting the lower affinity type III
inhibitors which have proven to be valuable starting points for
the development of more potent type II inhibitors.10,13 To
circumvent the difficulties associated with the use of activity
assays for the discovery of DFG-out binders, we set out to
develop a kinase binding assay using a type III inhibitor probe
combined with an enzyme fragment complementation techno-
logy (Figure 1a).14,15 The assay makes use of a labeled probe
consisting of a type III inhibitor chemically conjugated to a
small (∼5 kDa) enzyme donor (EDa) peptide fragment of
β-galactosidase. This conjugated enzyme fragment comple-
ments an inactive fragment of β-galactosidase (enzyme accep-
tor or EA), forming an active enzyme that generates a
Protein kinases serve as major factors in regulating cell
signaling pathways, since they are responsible for the transfer
of the γ-phosphate from ATP to various substrate proteins,
thereby modulating their function. Kinase disregulation can
contribute to diseases such as cancer, Alzheimer’s disease,
diabetes, and chronic inflammation.1-5 Over the past two
decades, protein kinases have become one of the major classes
of targets for drug discovery, resulting in the development of
multiple clinically approved kinase inhibitors.5,6 However,
limited selectivity of ATP-competitive (type I) inhibitors has
increasingly hampered drug development, particularly for
treatment of chronic diseases. Type I inhibitors bind to the
active state (DFG-in) of the kinase and have to compete with
high cellular concentrations of ATP. Additionally, the high
conservation of amino acids lining the ATP binding site
among all kinases makes the development of selective small
molecules particularly challenging. These impediments have
led to increased interest in identifying inhibitors with alter-
native binding modes that exploit the conformational varia-
bility of protein kinases.7 A major conformational change
undergone by many protein kinases is the switch between the
active DFG-in and inactive DFG-out conformation, during
which the rearrangement of the activation loop opens up
a hydrophobic pocket adjacent to the catalytic site.8
This allosteric pocket can be targeted by type III inhibitors
which stabilize the DFG-out state and prevent the kinase
domain from adopting the enzymatically competent DFG-in
†Atomic coordinates and structure factors for crystal structures of
compounds 4, 12, 19, 21, and 22 complexed with p38R can be accessed
using PDB codes 3HV7, 3HV6, 3HV3, 3HV4, and 3HV5, respectively.
*To whom correspondence should be addressed. Phone: þ49 (0)231
9742 6480. Fax: þ49 (0)231 9742 6479. E-mail: daniel.rauh@cgc.mpg.de.
aAbbreviations: ED, enzyme donor; EFC, enzyme fragment comple-
mentation; FRET, fluorescence resonance energy transfer; FLiK, fluor-
escent labels in kinases; Kd, dissociation constant; JNK2, c-Jun
N-terminal kinase 2; cpd, compound.
r
2009 American Chemical Society
Published on Web 11/23/2009
pubs.acs.org/jmc

358 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 1
Klu€ter et al.
Figure 1. Assay design and development. (a) Schematic representation of the structure-based assay principle (PDB code 1DPO).44 The enzyme
donor (ED)-ligand is bound to the allosteric site of the inactive kinase (DFG-out conformation). In a screening scenario, small molecules
compete with the ED-ligand for binding to the kinase. The peptide portion of the displaced ED-ligand binds to a truncated form of
β-galactosidase which is only active when complemented by the ED-ligand. The enzymatic activity of β-galactosidase is measured via
luminescence as a result of ED-ligand displacement. (b) Evaluation of pyrazoloureas as potential allosteric probes for p38R. Kd values were
obtained using a direct binding assay described previously.23,24 Because of its high affinity to p38R, 4 was selected to serve as the inhibitor part
of the ED-ligand. (c) Crystal structure of p38R in complex with 4. The inhibitor is bound to the allosteric site of inactive p38R. The 2Fo - Fc
electron density maps of the ligand (red) and the protein (gray) are contoured at 1σ. Hydrogen bonding interactions of the inhibitor with helix C
(blue) and the DFG-motif (orange) are shown as red dotted lines. (d) Surface representation of the allosteric binding site of p38R (gray) with
bound ligand (gray mesh) revealing the correct positioning of the amino group (blue sphere) to serve as an anchor point for further attachment
of the β-galactosidase peptide (red arrow). (e) Chemical structure of compound 4 and schematic representation of the ED-ligand probe. For
assay validation two molecules of compound 4 were linked to the ED peptide (see Schemes 1 and 2).
Scheme 1. Acetylation, Hydrolysis, and Maleimide Linker Synthesis of 4 (Type III Inhibitor)
chemiluminescent read-out signal. In the absence of displacing
ligands, it does not allow EFC with β-galactosidase to occur,
thereby resulting in minimum background signal. Competitive
displacement of the ED conjugate from the kinase occurs in the
presence of compounds which bind to the allosteric site,
allowing it to freely complement the EA and generate a
chemiluminescence signal as the assay readout.15,16 The ligand
portion of the ED probe, which has the ability to bind within
the allosteric pocket of the kinase, was first crystallized with the
target. The protein X-ray structure was solved in order to verify
that the ED peptide conjugate binds to the inactive state of
the kinase without causing any unexpected steric clashes
(Figure 1c,d). The crystal structure revealed the ideal point of
attachment for the conjugation of the β-galactosidase ED
peptide fragment to the ligand. This ED conjugate was then
validated against p38Rkinase using a series of known inhibitors
in conjunction with additional SAR studies and validation
of ligand binding modes by protein X-ray crystallography.
To get an idea about the applicability of this system to other
kinases, the type III ligand portion of the ED probe was

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 1 359
Scheme 2. Chemical Conjugation of ED Peptide with Type III Inhibitor
profiled against 58 kinases in a direct binding assay.17,18 The
results of this screen allowed the assay system to be extended to
additional kinases for screening of inhibitors that stabilize
inactive kinase conformations. Assay performance using
enzymatically active and inactive p38R (phosphorylated and
unphosphorylated) and the effect of added glycerol and DMSO
were examined.
probe to screen kinase inhibitors employing EFC technology
(Figure 1e). 4 was conjugated to the ED peptide by a linker to
yield the EFC probe, or ED-ligand, 4d (Schemes 1 and 2).
This ED-type III inhibitor conjugate was subsequently used
for the validation of the assay.
Validation of the Enzyme Fragment Complementation
(EFC) Assay System. Before using this system for IC50
determination of competing ligands, we set up and validated
the assay to ensure reliability and reproducibility. In general,
the assay requires two steps: (i) an incubation step in which the
target kinase, ED probe, and inhibitor are preincubated
and (ii) a detection step in which the enzyme acceptor (EA)
portion of β-galactosidase and its substrate are added to the
mixture such that the generated chemiluminescence signal
can be measured with time. The reproducibility and robust-
ness of the assay, measured as the output of the reaction step,
require that the kinase and ED-ligand have reached binding
equilibrium during the incubation step. Furthermore, we
titrated different concentrations of active p38R (activated via
phosphorylation by the upstream kinase MKK6) and inac-
tive p38R kinase (unphosphorylated) against the ED-ligand
and verified that the luminescence signal is proportional to
the ED-ligand-kinase complex (Figure 2a). Therefore, the
luminescence signal observed in a screening scenario would
be proportional only to the amount of ED-ligand displaced
by an added inhibitor. We found that 12.5 and 75 nM of
inactive and active p38R, respectively, are the highest con-
centrations of kinase that can be used to stay within the linear
range of the luminescence signal. These concentrations were
used for IC50 determinations, since they yielded the largest
dynamic range in signal change. The differences between
these concentrations could result from the fact that inactive
p38R is more likely to be in the DFG-out conformation,
which is preferred by the ED-ligand, thus requiring 6-fold
less kinase to scavenge up the ED-ligand. In the case of the
active kinase, a conformational change in the activation loop
must first occur to achieve the DFG-out conformation prior
to binding of the ED-ligand. This preference for the inactive
kinase was confirmed by measuring the binding constant
(Kd) of the ED-ligand toward both active and inactive p38R
(inactive p38R Kd = 3 ( 0.75 nM; active p38R Kd = 72 (
20 nM). The fact that type III inhibitors show different
Results
Design of the Type III Inhibitor Probe for ED Conjugation.
On the basis of the wealth of structural information available
for several potent types II and III inhibitors in complex with
p38R in the DFG-out conformation,19-22 we chose the weak
binder 1, which was the lead compound used to develop the
potent inhibitor 11 (BIRB-796),19 as a model to design and
synthesize three potent DFG-out inhibitors (2, 3, 4)
(Figure 1b) carrying a meta aniline at the N1 position of
the pyrazole moiety. On the basis of modeling studies, we
predicted that this amine would point toward the solvent and
would serve as a good anchor point for conjugation to the
ED peptide without compromising the binding affinity of the
molecule. To identify the best binder to use as a probe, we
made use of a novel fluorescent-labeled kinase binding assay
(FLiK) to directly determine the Kd of these allosteric
ligands.23 This assay confirmed that the measured com-
pounds bind to and stabilize the inactive conformation of
the kinase. We found that the naphthyl derivative 4 has the
highest affinity for p38R kinase with a Kd of ∼12 nM
(Figure 1b).
Crystal Structure of a Type III Inhibitor with p38r. To
verify that 4 adopts the proposed binding mode in the
allosteric pocket with the newly introduced amino group
pointing out of the binding site, we solved the crystal
structure of this compound in complex with p38R
(Figure 1c) (Table 1). 4 binds to the DFG-out conformation
and resides in the allosteric site, as previously seen in the
complex of the same compound with cSrc (PDB code
3F3T).24 As expected from our modeling studies with
p38R, the free anilino moiety points into the solvent and
qualifies as an anchor point for further conjugation/modifi-
cations (Figure 1d). Thus, we chose this inhibitor to con-
jugate to the enzyme donor (ED) peptide and used this as the

360 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 1
Klu€ter et al.
Figure 2. Validation of the assay system in p38R. (a) Increasing amounts of active (white circles) and inactive (black circles) kinase were
titrated against 1.25 nM ED-ligand. The percentage of ED-ligand bound was calculated from the β-galactosidase signal representing the
amount of free ligand in solution. (b, c) Dose response curves of 4 in the presence of DMSO and glycerol, respectively.
affinities to active and inactive kinases has been reported
previously.25,26 With the focus on stabilizers of the inactive
conformation, inactive kinase was then used for further
optimization and IC50 determination. We also validated
the additive tolerance of the assay by measuring the effect
of DMSO and glycerol concentration. As shown in
Figure 2b, a maximum of 2.5% DMSO (v/v) could be used
in the assay, since higher amounts tend to lower the dynamic
range (i.e., signal/noise ratio). A plausible explanation for
the drop in the assay window with higher concentrations of
DMSO could be the compromised stability of the kinase
during prolonged incubation times. Similarly, glycerol at
concentrations >1% (v/v) significantly lowers the dynamic
range of the assay system (Figure 2c). To maintain the
robustness of the assay, it is therefore important to minimize
the amount of glycerol (a component found in many kinase
storage buffers) in the final mixture. To attain equilibrium of
the ED probe with the kinase, we titrated the ED-ligand
against the enzyme acceptor (EA) to further optimize the
assay response and to minimize background signal. We
found that the amount of EA had little influence on the
signal-to-noise ratio. With respect to different ED-ligand
concentrations, the maximal signal-to-noise ratio was ob-
tained using 4 nM ED-ligand together with equal amounts of
the EA portion of β-galactosidase and chemiluminescence
substrate solution. However, in order to increase the like-
lihood that the competing inhibitor added to the reaction
mixture would successfully compete out the ED-ligand from
the target kinase, we chose only 1.25 nM ED-ligand at a
small expense to the dynamic range (i.e., maximal signal is
reduced when less ED-ligand is used). Having established
the ideal concentrations of the various assay components,
we set out to determine the shortest incubation and detec-
tion times needed to yield the highest signal. We found that
the incubation time for the kinase and the ED-ligand
had little effect on the signal-to-noise ratio of the readout
and used a 2 h incubation time for further experiments.
Although the detection step itself was shown to run linearly
for at least 6 h, we chose a detection time of 1 h for
convenience.
Characterization of Kinase Inhibitors. After optimizing the
assay conditions, we turned our attention to the measure-
ment of ED-ligand displacement by some known kinase
inhibitors in order to further validate the assay response.
Titration with unlabeled 4 yielded a sigmoidal dose response
curve with IC50 values in the nanomolar range (Figure 2b,c).
This indicated that the ED conjugate can get displaced
by suitable inhibitors and that the binding of ED probe
to this site is specific and the ED probe can be used to identify
and characterize small molecular binders of kinases. More
importantly, the IC50 value is consistent with the results
obtained using an activity-based HTRF (homogeneous
time-resolved fluorescence) kinase assay. Since there did
not appear to be any dependence of the IC50 on the pre-
incubation (reaction) step, we used a 2 h incubation as
standard for measuring the IC50 of further compounds.

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 1 361
Table 1. Data Collection and Refinement Statistics for p38R with 4, 12, 19, 21, and 22
p38R with 4 (3HV7)
P212121
p38R with 12 (3HV6) p38R with 19 (3HV3) p38R with 21 (3HV4) p38R with 22 (3HV5)
space group
cell dimensions
P212121
P212121
P21
P21
˚
a, b, c (A)
66.71, 74.52, 78.91
90.00, 90.00, 90.00
40.0-2.40
66.41, 74.66, 78.35
90.00, 90.00, 90.00
40.0-1.95
66.51, 74.52, 76.98
90.00, 90.00, 90.00
40.0-2.00
70.64, 74.06, 72.00 70.77, 74.12, 71.60
90.00, 93.90, 90.00 90.00, 93.72, 90.00
R, β, γ (deg)
˚
resolution (A)
40.0-2.60
(2.70-2.60)a
11.1 (49.9)
19.1 (5.4)
99.5 (99.3)
3.5 (3.5)
40.0-2.25
(2.35-2.25)a
4.6 (25.0)
19.8 (4.5)
99.0 (99.3)
2.6 (2.6)
(2.50-2.40)a
6.9 (35.7)
17.7 (5.2)
(2.05-1.95)a
3.3 (21.7)
19.1 (7.7)
(2.10-2.00)a
4.3 (37.5)
19.8 (4.9)
Rsym or Rmerge (%)
I/σI
completeness (%)
redundancy
refinement
99.9 (100)
4.8 (4.8)
99.7 (99.9)
4.7 (4.6)
98.8 (98.8)
4.8 (4.9)
˚
resolution (A)
no. reflections
40.0-2.40
15869
20.3/27.9
40.0-1.95
28963
20.9/25.4
40.0-2.00
26160
23.0/30.0
40.0-2.60
22868
19.4/27.1
40.0-2.25
34893
18.8/26.0
R
work/Rfree
no. atoms
protein
2661
50
2668
70
2718
64
5434
148
5453
152
ligand/ion
water
107
177
124
157
295
B-factors
protein
27.0
27.2
17.6
25.7
29.9
29.6
33.4
33.2
41.0
40.9
41.3
43.0
17.6
17.5
23.3
13.7
28.2
28.2
29.0
27.0
ligand/ion
water
rms deviations
˚
bond lengths (A)
0.016
1.620
1.54170
100
0.015
1.503
1.54170
100
0.016
1.542
1.00000
90
0.014
1.598
0.97703
90
0.016
1.613
bond angles (deg)
˚
wavelength (A)
temperature (K)
X-ray source
1.54170
100
Bruker AXS Microstar
Bruker AXS Microstar Bruker AXS Microstar SLS
SLS
Ramachandran plot
residues in
most favored regions (%) 87.8
90.2
9.5
89.0
10.3
85.9
13.5
89.0
10.8
additional allowed
regions (%)
11.2
generously allowed
regions (%)
disallowed regions (%)
1.0
0.3
0.7
0.7
0.2
0.0
0.0
0.0
0.0
0.0
a For all p38R complex structures, diffraction data from one crystal were used to determine the structure. Values in parentheses are for the highest
resolution shell.
Using these optimized conditions, we determined IC50
values for a set of 30 compounds chosen to represent
the three kinase inhibitor types known for p38R (Chart 1,
Table 2). Staurosporine (30),18 which is a nonbinder of
p38R, served as a negative control and as expected did not
displace the ED-ligand. 10 was also used as a negative
control because its 2,6-dicloro-3-trifluormethylphenyl
moiety does not fit into the allosteric pocket because
of steric hindrance.23 No interference of β-galactosidase
enzyme activity was detected for any of the compounds
tested in the EFC kinase binding assay (data not shown). We
observed mean Z0-factors of 0.7-0.8. The Z0-factor is a
measure of the statistical quality or power of a high-throughput
screening (HTS) assay,27 and the Z0-factors generated by the
EFC assay are more than suitable for screening large com-
pound libraries.
X-ray Crystallographic Analysis. To better understand the
affinity profiles of all tested compounds, we solved the
crystal structures of quinolines 19, 21, and 22 as well as of
12, an analogue of 11, in complex with p38R (Figure 3)
(Table 2). 12 was chosen to help explain its 10-fold lower
binding potency when compared to 11, a closely related
derivative and a well-known type II p38R inhibitor20
(Figure 3). Regan and co-workers recently reported a lower
potency of 12 when compared to 11 and argued that the loss
of hydrophobic interactions observed when the naphthyl
moiety of 11 interacts with the lipophilic subpocket behind
the gatekeeper residue Thr106 could result in an unfavorable
orientation of the ethoxymorpholino unit of 12, preventing
an effective hinge region contact.22 However, superimposi-
tion of p38R-12 with the p38R-11 complex (PDB code
1KV2) revealed that both ligands align almost perfectly,
adopting the same binding site geometries including a hydro-
gen bond to the backbone of Met109 of the hinge region
(Figure 3a). Therefore, the significant increase in binding
affinity observed for 11 is solely due to its naphthyl moiety
which, compared to the phenyl ring of 12, limits confor-
mational flexibility and penetrates much deeper into the
hydrophobic backpocket, thereby substantially increasing
binding affinities to p38R. This further highlights the
importance of this subpocket for ligand binding in p38R.
We observed a similar trend when comparing the binding
characteristics of the 1,4-fused quinoline 19 and the 1,3-fused
quinoline 21, respectively (Figure 3b). In both structures, the
quinoline moiety of the ligand binds via a hydrogen bond to the
hinge region and adopts a binding mode isostructural to
quinazoline-based type II inhibitors of p38R.13,23 However,
the phenyl moiety that connects the hinge region binding
portion of the ligands with the pyrazolourea portion clearly
shows that the 1,3-fused molecule 21 accommodates the
hydrophobic pocket much better than the 1,4-fused com-
pound 19 because of its more favorable binding geometry.

362 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 1
Chart 1. Type I, II, and III Inhibitors
Klu€ter et al.
Table 2. IC50 Values of Type I, Type II, and Type III Inhibitors (Chart 1) for p38R and JNK2a
p38R kinase: IC50 (nM)
p38R kinase: IC50 (nM)
JNK2 kinase: IC50 (nM)
cpd
EFC
HTRF
cpd
EFC
HTRF
cpd
EFC
HTRF
90 ( 35
1300 ( 470
150 ( 60
12600 ( 4300
570 ( 350
150 ( 30
1400 ( 500
380 ( 170
76 ( 25
1
16800 ( 7400
440 ( 40
1100 ( 60
170 ( 30
220 ( 40
220 ( 170
45 ( 25
21000 ( 8000
1100 ( 260
2000 ( 900
90 ( 40
100 ( 6
120 ( 20
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
1100 ( 370
420 ( 200
370 ( 30
470 ( 70
420 ( 130
310 ( 50
95 ( 25
230 ( 30
250 ( 90
400 ( 140
96 ( 50
1500 ( 650
320 ( 20
460 ( 320
900 ( 420
2400 ( 60
290 ( 80
150 ( 40
290 ( 160
140 ( 30
1100 ( 710
36 ( 7
4
70 ( 25
440 ( 110
27 ( 7
1800 ( 540
770 ( 15
1200 ( 70
1900 ( 820
100 ( 15
220 ( 60
2
8
3
11
12
14
17
18
19
27
4
5
6
7
150 ( 80
8
1300 ( 180
680 ( 60
nb
1200 ( 260
1100 ( 360
nb
9
10
11
12
13
14
15
220 ( 170
2300 ( 260
760 ( 450
1400 ( 530
1000 ( 580
250 ( 30
5500 ( 2800
1800 ( 1400
800 ( 220
6000 ( 1400
75 ( 25
35 ( 8
820 ( 150
1000 ( 150
nb
110 ( 40
2800 ( 520
nb
a The structures of type I, type II, and type III inhibitors are shown in Chart 1. On the basis of their binding mode to the kinase domain of p38R, the
individual inhibitors are classified as follows: 1-12, type III; 13-25, type II; 26-30, type I. IC50 values were obtained using the developed EFC
displacement assay (enzymatically inactive p38R and JNK2 (unphosphorylated)). For comparison, IC50 values were measured with an activity based
phosphorylation assay (HTRF) (enzymatically active p38R and JNK2 (phosphorylated)). Both assay systems give comparable and consistent IC50
values, demonstrating that the displacement assay can be used for inhibitor screening.
Occupation of this pocket appears to be animportant feature of
these hybrid compounds and explains the overall better binding
potency of the 1,3-quinoline hybrids compared to their 1,4-
substituted counterparts. The same trend is conserved among

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 1 363
Figure 3. Crystal structures of various pyrazoloureas in complex with p38R. Colored areas in (a) and (b) represent different binding loci in and
around the ATP cleft typically addressed by pyrazolourea-based type II inhibitors. The hinge region of the kinase is in pink. (a) Alignment of 12
(gray) with 11 (green) bound to p38R revealing the importance of the hydrophobic subpocket for ligand binding. (b) Overlay of the cocrystal
structures of p38R with the 1,3-fused quinoline 21 (gray) and 1,4-fused quinoline 19 (green). 21 addresses the hydrophobic subpocket much
better than 19, resulting in a higher binding affinity. (c) Structure of the quinoline 22 bound to p38R. The 2Fo - Fc electron density maps of the
ligand (red) and the protein (gray) are contoured at 1σ. (d) Overlay of 22 (gray) with the corresponding amine analogue 21 (yellow) in complex
with p38R. Both molecules align well, and the quinoline cores form π-π interactions with the side chain of Phe169 of the DFG-motif. The free
amine of 21 and the nitro group of 22 in the 6-position of the quinoline core are in proximity to the backbone of Val30 of the glycine-rich loop.
quinazoline-based hybrid compounds as well (compare 14
and 16 to 17 and 18).
Selectivity, Profiling, and Extension of EFC Based Kinase
Binding Assay to Additional Kinases. In order to determine
selectivity profiles of the pyrazolourea 4 and to identify addi-
tional target kinases that (i) can adopt the DFG-out conforma-
tion and (ii) would be most likely suitable for the enzyme
fragment complementation (EFC) assay, kinase inhibitor
profiling was performed against a selected subpanel of 58
recombinantly expressed kinases17 at 10 μM (Figure 4a). The
inhibitor profile of 4 shows strong binding to surprisingly
distinct kinases which most likely can adopt the DFG-out
conformation. This finding can be explained on the basis of
the comparison of crystal structures of 4 in complex with p38R
and cSrc.24 The naphthyl moiety of the ligand binds in a
hydrophobic subpocket close to the gatekeeper residue. The
differently sized gatekeeper side chains control access to the
allosteric site from the direction of the hinge region and thereby
serve as a critical determinant of inhibitor selectivity and affi-
nity.28,29 Facilitated by our ED-ligand probe, which is based on
the type III binding mode 4, we were able to adapt this assay
format to additional kinases such as JNK2, DDR2, PDGFR-A,
and RET (Figure 4b).
Another interesting result of the structure-activity rela-
tionship studies (SAR) was the overall better binding
potency of compounds that contain a nitro group in the
6-position of the quinazoline or quinoline inhibitor core
when compared to their amino-substituted counterparts.
To better understand this trend, we solved the crystal struc-
ture of the 6-nitro substituted quinoline 22 in complex with
p38R and compared its binding mode with the amino sub-
stituted analogue 21 (Figure 3c,d). The ligand is well-defined
by its electron density and adopts a binding mode typically
observed for this class of type II inhibitors.23 Superimposing
both structures showed almost identical binding modes and
gave no clear explanation for the increased binding affinity
of 22 compared to 21 (Figure 3d). It is likely, however, that
the different substituents in the 6-position of the quinazoline
core (nitro versus amine) affect the strength of the π-π
interactions with the nearby side chain of Phe169 of the DFG
motif and also influence the basicity of the quinoline nitro-
gen which forms the crucial hydrogen bond to the backbone
of the hinge region. In summary, the presence of a nitro
group on the quinoline core in combination with the 1,3
substitution pattern (deeper binding into the hydrophobic
subpocket) may help to explain why 22 shows the highest
affinity toward p38R of all tested hybrid compounds.
JNK2 Kinase Binding Assay. Out of the four additional
kinases tested, we picked JNK2 for further validation and
screened a selection of pyrazoloureas. Although the ED-
ligand probe based on 4 has a lower affinity for JNK2, we
were able to reliably study the SAR of several compounds

364 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 1
Klu€ter et al.
Figure 4. Kinase selectivity profile of a type III inhibitor and evaluation of the applicability of the EFC assay to other kinases. (a) Panel screen
of the allosteric kinase inhibitor 4 at 10 μM against a subset of 58 human kinases employing the KINOMEscan binding assay (Ambit
Biosciences). Weak or no binding is represented by a high number; tight binding is indicated by low numbers. A KINOMEscan score threshold
of 30 (dashed lines) is used to dissect binders from nonbinders. (b) Dose response curves of 4 titrated against a panel of four kinases in the EFC
assay demonstrate that the principle using an allosteric probe as a tool for drug discovery can be adapted to other kinases as well.
(Table 2). In contrast to what was observed in p38R, we
found that the 1,4-fused hybrid compounds 14 and 19 are
more potent when compared to their 1,3-fused counterparts
17 and 18. In JNK2 kinase, access to the hydrophobic
subpocket is regulated by a bulkier methionine (Met108)30
gatekeeper while p38R carries a smaller amino acid (Thr106)
at this position. In crystal structures of wild type cSrc and the
drug resistant mutant variant cSrc-T338M in complex with
14, we have recently observed that the central 1,4-fused
phenyl moiety of the inhibitor avoids sterically demanding
gatekeeper side chains via free rotation along the 1,4-axis.13
Such observed flexibility likely explains the preference for
these inhibitors with JNK2 kinase.
conformation becomes more limited to detection and may
result in an enrichment of hits of the type I scaffold in an HTS
screening scenario. To address these needs, we developed and
characterized a kinase binding assay based on enzyme frag-
ment complementation technology which allows for the iden-
tification of small molecule stabilizers of inactive kinase
conformations. To directly target a site outside the highly
conserved ATP cleft, we used principles of structure-based
ligand design to develop and synthesize type III inhibitors that
would bind to the allosteric site of inactive (DFG-out) kinase
conformations and allow for conjugation with an ED peptide
fragment. The cocrystal structure of p38R in complex with
compound 4 ultimately confirmed the precise binding mode
and proved that the anilino moiety in the 3-position of the
phenyl ring points out into the solvent and serves as a good
anchor point for further attachments of the ED peptide using
different linkers. We showed that displacement of the type III
inhibitor-ED peptide conjugate by suitable binders leads to
an increased signal, allowing a luminescence reaction to serve
as the readout for the assay system. Facilitated by this assay
system, we then determined IC50 values for various type II
and III ligands with high reproducibility and robustness.
The setup of the assay is simple and straightforward and,
most importantly, does notrely on enzymaticallyactivekinase
preparations, which is crucial for the enrichment of hits which
bind to inactive kinase conformations. We feel that the
detection system based on enzyme fragment complementa-
tion and chemiluminescence to identify stabilizers of the
inactive kinase conformation offers major advantages over
other fluorescence polarization-based approaches. We show
Discussion and Conclusion
Kinase inhibitor research has dramatically progressed in
the past decade and has resulted in a variety of approved drugs
for targeted therapies. However, several challenges such as the
limited inhibitor selectivity of classical hinge region binders
(type I) and acquired mutation-associated drug resistance
remain to be solved. Ligands that address alternative binding
pockets and thereby stabilize inactive kinase conformations
are thought to offer new opportunities to circumvent these
limitations. State of the art kinase assays rely onenzymatically
active kinases and measure the perturbation of substrate
phosphorylation as a means of target inhibition. Such formats
require the kinase to be phosphorylated which shifts the
equilibrium from the inactive DFG-out toward the enzyma-
tically competent DFG-in conformation. As a direct con-
sequence, binding of type II/III inhibitors to the DFG-out

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 1 365
superior Z0-factors for our assay demonstrating its high
sensitivity, which is often less optimal in fluorescence polar-
ization experiments. The biggest advantage deals with the
intrinsic fluorescence of inhibitors, which can be quite high in
some small molecules, and often interferes with fluorescence-
based approaches, thereby impairing the screening results and
the detection of false hits. Lastly, we showed that the assay
technique presented here can be applied to other kinases that
can adopt the desired DFG-out conformation. Furthermore,
the simple mix-and-ready approach of the assay, the mini-
aturization of the assay volume to 20 μL, and the good
Z0-factors demonstrate that the kinase binding EFC assay
can be employed to screen for type II and type III inhibitors as
stabilizers of inactive kinase conformations in an HTS com-
patible format.
solution. The reaction mixture was stirred at ambient tempera-
ture overnight. The solvent was concentrated under reduced
pressure on a rotary evaporator and the residual oil subjected to
purification using RP-HPLC to afford 1.0 mg of pure product.
ESI-MS: [Mþ] = 580 (found); 580 (calcd).
Reduction of Linear Enzyme Donor (ED) Peptide. Enzyme
donor (ED) peptide (5 mg) was dissolved in water (1 mL) and
treated with DTE (5 mg) overnight at ambient conditions
followed by RP-HPLC purification (MeCN/H2O with 0.1%
TFA) to yield pure 2.5 mg (50%) linear ED having free
sulfhydryls. This was then immediately employed in the next
step for conjugation with 4c.
Conjugation of ED Peptide with Compound 4c (4d). To the
above freshly purified linear ED (2.8 mL, 100 μM), 100 mM
sodium phosphate buffer, pH 8.0 (200 μL), was added to adjust
the pH to be between 6.3 and 6.7. To this, water (1 mL) and
DMF (100 μL) were added, and 10-fold excess of maleimide
linked compound 4c in DMF/H2O (100 μL) was added dropwise
under vigorous stirring for 5 min. The mixture was left at
ambient conditions for 90 min, and then an aliquot was taken
to monitor the reaction mixture on RP-HPLC and to ensure
complete reaction of the ED peptide sulfhydryls. The ED-ligand
conjugate was purified to homogeneity by RP-HPLC, and the
pure fractions collected were lyophilized and resuspended in
MeCN/H2O mixture. The concentration of the enzyme donor
(ED) inhibitor peptide conjugate was assessed by UV-vis
spectroscopy with an extinction coefficient calculated to be
36,300 at 280 nm (∼15-20% unoptimized yield). MALDI-
MS: [Mþ] = 6549 (found); 6549 (calcd).
Materials. All reagents, fine chemicals, and solvents were
purchased from Acros, Fluka, Sigma, Aldrich, or Merck, unless
otherwise stated in the text. Crystallization plates (Easy Xtal
Tool; 24-well) were obtained from Qiagen GmbH (Hilden,
Germany). White microtiter plates in a 384-well small volume
format used for EFC assay and HTRF assay were obtained
from Greiner Bio-One GmbH (Solingen, Germany). Reagents
for the activity-based assay were purchased from Cisbio
Experimental Section
Chemistry. Synthesis and characterization of compounds
1-24 have been described previously.13,23,24 25-29 were kindly
provided by Merck-Serono (Darmstadt, Germany). 30 was
purchased from Merck Chemicals (Darmstadt, Germany). All
compounds have a purity of g95% as determined by HPLC
coupled with mass spectrometry (HPLC-ESI-MS). HPLC-
ESI-MS analyses were performed on an HPLC system from
Agilent (1200 series) with an Eclipse XDB-C18 5 μm (column
dimensions: 150 mm ꢀ 4.60 mm) column from Agilent and UV
detection at 254 nm coupled to a Thermo Finnigan LCQ
Advantage Max ESI spectrometer. Solvent system was H2O
with 0.1% formic acid (solvent A) and acetonitrile with 0.1%
formic acid (solvent B) and was used at a flow of 1 mL/min.
Gradient was a follows: 0 min, 90% solvent A/10% solvent B;
1 min, 90% solvent A/10% solvent B; 10 min, 0% solvent
A/100% solvent B; 12 min, 0% solvent A/100% solvent B;
15 min, 90% solvent A/10% solvent B.
Synthesis of the Enzyme Donor (ED) Conjugate for Compound
4. The synthesis of the ED-inhibitor probe involves three steps
(acetylation of 4 to 4a, hydrolysis to 4b, and maleimide linker
coupling to 4c) prior to ED conjugation to yield 4d. Step by step
procedures are outlined below.
ꢀ
(Bagnols-sur-Ceze, France). All kinases, except for p38R, were
obtained from Invitrogen (Carlsbad, CA). For synthesis of the
enzyme donor conjugate of 4, RP HPLC was performed on a
Waters 600 equipped with a PDA detector and a Zorbax 300A
C18 semiprep column. Other reagents such as enzyme acceptor
(EA) enzyme core assay buffer, chemiluminescent (CL) sub-
strate, and enzyme donor dilution buffer (EDDB) were from
DiscoveRx, Fremont, CA. The enzyme donor (ED) peptide was
synthesized by American Peptide, Sunnyvale, CA.
Methyl
2-(3-(3-tert-Butyl-5-(3-naphthalen-1-ylureido)-1H-
pyrazol-1-yl)phenylamino) acetate (4a). 4 (5 mg, 1.9 mmol) was
suspended in 0.5 mL of anhydrous DMF followed by addition
of diisopropylethylamine (2 equiv) and bromethyl acetate
(1.1 equiv), and the mixture was stirred at ambient temperature
for 4-5 h. The reaction was monitored on TLC and RP HPLC.
The DMF was removed in vacuo, and the concentrated oil was
subjected to purification by RP-HPLC (MeCN/H2O with 0.1%
TFA). The expected product was isolated and lyophilized to
afford 3.7 mg (78%) of an amorphous powder. ESI-MS: [M þ
Hþ] = 473 (found); 472 (calcd).
2-(3-(3-tert-Butyl-5-(3-naphthalen-1-ylureido)-1H-pyrazol-1-yl)-
phenylamino) acetic Acid (4b). A solution of 4a (0.30 g, 1.78
mmol) was subjected to alkaline hydrolysis using MeOH-water
and 2 N NaOH (100 μL) with the pH maintained at 9.0 at
ambient temperature for 1 h. The reaction was monitored on RP
HPLC. The solvent was concentrated under reduced pressure
and the residue purified by RP HPLC (MeCN/H2O with 0.1%
TFA). The pure fraction was lyophilized to afford 1.8 mg (40%)
of the pure product. ESI-MS: [M - Hþ] = 456 (found); 457
(calcd).
2-(3-(3-tert-Butyl-5-(3-naphthalen-1-ylureido)-1H-pyrazol-1-yl)-
phenylamino)-N-(2-(2,5-dioxo-2,5-dihydro-1H-pyrrol-1-yl)ethyl)-
acetamide (4c). A solution of derivatized compound 4b (1.5 mg,
0.68 mmol) was dissolved in 0.5 mL of anhydrous DMF and
cooled at 0 °C followed by the addition of HBTU (1.2 mg, 0.45
mmol). N-(2-Aminoethyl)maleimide TFA (1.92 g, 0.3 mmol)
freshly prepared in anhydrous DMF (100 μL) was neutralized
with N-methylmorpholine (2 μL) and added to the above
Kinase Expression and Purification. The p38R construct
was cloned into a pNFG vector and was transformed as an
N-terminal His-tag construct with thrombin cleavage site into
BL21(DE3) Rosetta E. coli. Cultures were grown at 37 °C until
an OD600 of 0.5 was obtained, cooled to 18 °C using an ice bath,
and then induced with 0.5 mM IPTG for overnight (∼20 h)
expression at 18 °C while shaking at 160 rpm. Cells were lysed in
buffer A (50 mM Tris, pH 8.0, 500 mM NaCl þ 5% glycerol þ 25
mM imidazole) and loaded onto a 30 mL Ni column (self-
packed), washed with 3 column volumes (CV) of Ni buffer A,
and then eluted with a 0-50% linear gradient using Ni buffer B
(Ni buffer A þ 500 mM imidazole) over 2 CV. The protein was
dialyzed for 5 h against dialysis buffer (25 mM HEPES, pH 7.0,
5% glycerol, 100 mM NaCl, 1 mM DTT) and then cleaved by
incubating with thrombin protease (0.2 U/mg uncleaved
protein) in a 50 mL falcon tube overnight at 4 °C. The protein
was then centrifuged for 15 min at ∼13 000 rpm to remove any
precipitate that may have formed during the cleavage step.
The supernatant was then taken and diluted at least 4-fold in
anion buffer A (50 mM Tris, pH 7.4, 5% glycerol, 50 mM NaCl,
1 mM DTT) and loaded onto a 1 mL Sepharose Q FF column
(GE Healthcare) and washed with 10 CV of anion buffer A.
The protein was eluted with a 0-100% linear gradient of
anion buffer B (anion buffer A þ 600 mM NaCl) over 20 CV.

366 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 1
Klu€ter et al.
The protein was pooled and concentrated to 2 mL and passed
through a Sephadex HiLoad 26/60 Superdex 75 column equili-
brated with size exclusion buffer (20 mM HEPES, pH 7.1, 5%
glycerol, 50 mM NaCl, 100 mg/L methionine, 10 mM DTT) at a
rate of 2 mL/min. The eluted protein was then concentrated to
∼15 mg/mL, aliquoted, and frozen at -80 °C.
globulins) and added to 5 μL of ED-ligand in a final concentra-
tion of 1.25 nM to initiate the kinase incubation reaction. After
2 h of incubation, 5 μL of enzyme acceptor (EA) and 5 μL of the
chemiluminescence substrate (CL substrate) (both from Dis-
coveRx “Kinase Binding Assay Detection Kit”) were included
for the complementation reaction for 1 h or more before
measuring the plate.
IC50 determinations of inhibitors were carried out by prepar-
ing nine 50-fold inhibitor dilutions prepared in DMSO and
diluted in EFC assay buffer to have a 2% DMSO/buffer
solution. Then 2.5 μL of 4ꢀ kinase (12.5 nM final con-
centration) and 2.5 μL of 4ꢀ ED-ligand (1.25 nM final con-
centration) were incubated together with 5 μL of inhibitor/
buffer solution for 2 h prior to a 1 h complementation reaction.
The final DMSO concentration was 1%, and inhibitor concen-
trations ranged from 100 μM to 1.5 nM. Chemiluminescence
was measured with a Tecan Safire2 plate reader.
Dose Response Curve of Compound 4 on PDGFR B, DDR2,
RET, and JNK2. The same protocol as described for IC50
determination in p38R kinase was used except the reaction time
of β-galactosidase formation was only 10 min before measuring.
Kinases were used at a concentration of 75 nM for PDGFR B
(P3082, enzymatically active), DDR2 (PV3870, enzymatically
active), and RET (PV3819, enzymatically active), while 150 nM
was chosen for JNK2 (PV3621, enzymatically inactive).
Crystallization of p38r-Inhibitor Complexes. Various inhi-
bitors were cocrystallized with purified p38R using conditions
similar to those previously reported.31 Briefly, protein-inhibi-
tor complexes were prepared by mixing 40 μL of p38R (10-
12 mg/mL) with 0.4 μL of inhibitor (100 mM in DMSO) and
incubating the mixture for 1-2 h on ice. Crystals were grown in
24-well crystallization plates using the hanging drop vapor
diffusion method and by mixing 1.5 μL of protein-inhibitor
solution with 0.5 μL of reservoir (100 mM MES, pH 5.6-6.2,
20-30% PEG4000, and 50 mM n-octyl-β-glucopyranoside).
Structure Determination, Refinement, and Crystallographic
Statistics. For crystals of p38R with inhibitors, 20% glycerol
was added to the reservoir buffer as a cryoprotectant before
freezing in liquid nitrogen. Diffraction data of the p38R-19 and
p38R-21 complex crystals were collected at the PX10SA beam-
line of the Swiss Light Source (PSI, Villigen, Switzerland) using
˚
wavelengths close to 1 A. Diffraction data of the p38R-4,
p38R-12, and p38R-22 complexes were collected in-house.
All data sets were processed with XDS32 and scaled using
XSCALE.32 All p38R-inhibitor complex structures were solved
by molecular replacement with PHASER33 using the published
p38R structures (PDB code 1ZYJ)34 or (PDB code 2EWA)8 as
templates. The molecules in the asymmetric unit were manually
modified using the program COOT.35 The model was first
refined with CNS36 using simulated annealing to reduce model
bias. The final refinement was performed with REFMAC.37
Inhibitor topology files where generated using the Dundee
PRODRG2 server.38 Refined structures were validated with
PROCHECK.39 Data collection, structure refinement statistics,
PDB codes, and further details for the data collection as well
as Ramachandran plot results are shown below in Table 2.
PyMOL40 was used to produce the figures.
Acknowledgment. We thank Guido Zaman, Willem van
Otterlo, Beate Aust, Michael Beck, Oliver Gutbrod, Svend
Matthiesen, Martine Ercken for helpful discussions and Rajini
Bompelli for technical assistance. J.R.S. was funded by the
Alexander von Humboldt Foundation. Schering Plough,
Bayer-Schering Pharma, Merck-Serono, and BayerCrop
Science are thanked for financial support. The work was
supported by the German Federal Ministry for Education
and Research through the German National Genome Research
Network-Plus (NGFN-Plus) (Grant No. BMBF 01GS08102).
Activation of p38r Kinase. Active MKK6 (T207E, T211E)
(as purchased from Invitrogen) was diluted to 0.83 μg/mL
together with inactive p38R (3.33 μg/mL final concentration)
in buffer containing 50 mM Tris-HCl (pH 7.5), 1 mM Na3VO4,
0.01% NP-40, 0.1 mg/mL BSA, 2 mM DTT, 10 mM MgCl2,
1 mM EGTA and was incubated together with 200 μM ATP at
30 °C for 1 h with constant agitation. ATP was subsequently
removed by overnight dialysis in buffer containing 25 mM Tris,
pH 8.0, 150 mM NaCl, and 10% (v/v) glycerol.
IC50 Determination with a Biochemical in Vitro HTRF Assay.
IC50 determinations based on an enzyme activity assay were mea-
sured with the HTRF approach. GST-tagged ATF2 (200 nM) was
obtained from CST (Boston, MA) and used as a substrate, and after
45 min of incubation time, 4 ng/μL of an antiphosphotyrosine
antibody labeled with europium cryptate was added together with
12.5 nM of an anti-GST antibody labeled with the fluorophore D2.
The FRET between europium cryptate and D2 was measured to
quantify the phosphorylation of the ATF2. ATP and substrate
concentrations were set at the KM values, and kinases were used in
concentrations of 1 ng/μL for activated p38R and 0.65 ng/μL for
activated JNK2. A Safire2 plate reader (Tecan) was used to measure
the fluorescence of the samples at 620 nm (Eu-labeled antibody) and
665 nm (D2-labeled antibody) 60 μs after excitation at 317 nm. The
quotient of both intensities for reactions made with nine different
inhibitor concentrations was fit to a Hill four-parameter equation to
determine IC50 values. Each reaction was performed in duplicate or
triplicate on at least three independent occasions.
References
(1) Sandoval, D. A.; Obici, S.; Seeley, R. J. Targeting the CNS to treat
type 2 diabetes. Nat. Rev. Drug Discovery 2009, 8, 386–398.
(2) Cohen, P. Targeting protein kinases for the development of anti-
inflammatory drugs. Curr. Opin. Cell Biol. 2009, 21, 317–24.
(3) Fabbro, D.; Ruetz, S.; Buchdunger, E.; Cowan-Jacob, S. W.;
Fendrich, G.; Liebetanz, J.; Mestan, J.; O’Reilly, T.; Traxler, P.;
Chaudhuri, B.; Fretz, H.; Zimmermann, J.; Meyer, T.; Caravatti,
G.; Furet, P.; Manley, P. W. Protein kinases as targets for anti-
cancer agents: from inhibitors to useful drugs. Pharmacol. Ther.
2002, 93, 79–98.
(4) Gaestel, M.; Kotlyarov, A.; Kracht, M. Targeting innate immu-
nity protein kinase signalling in inflammation. Nat. Rev. Drug
Discovery 2009, 8, 480–499.
(5) Zhang, J.; Yang, P. L.; Gray, N. S. Targeting cancer with small
molecule kinase inhibitors. Nat. Rev. Cancer 2009, 9, 28–39.
(6) Janne, P. A.; Gray, N.; Settleman, J. Factors underlying sensitivity
of cancers to small-molecule kinase inhibitors. Nat. Rev. Drug
Discovery 2009, 8, 709–723.
(7) Bogoyevitch, M. A.; Fairlie, D. P. A new paradigm for protein
kinase inhibition: blocking phosphorylation without directly
targeting ATP binding. Drug Discovery Today 2007, 12, 622–633.
(8) Vogtherr, M.; Saxena, K.; Hoelder, S.; Grimme, S.; Betz, M.;
Schieborr, U.; Pescatore, B.; Robin, M.; Delarbre, L.; Langer, T.;
Wendt, K. U.; Schwalbe, H. NMR characterization of kinase p38
dynamics in free and ligand-bound forms. Angew. Chem., Int. Ed.
2006, 45, 993–997.
€
(9) Backes, A. C.; Zech, B.; Felber, B.; Klebl, B.; Muller, G. Small-
molecule inhibitors binding to protein kinase. Part II: the novel
pharmacophore approach of type II and type III inhibition. Expert
Opin. Drug Discovery 2008, 3, 1427–1449.
Optimization of the EFC Assay System and IC50 Determina-
tion. For the determination of a suitable kinase concentration,
5 μL of kinase at a given concentration was diluted in EFC
buffer (35 mM HEPES, pH 7.4, 20 mM NaCl, 0.4 mM EGTA,
0.01% Tween-20, 10 mM MgCl2, 0.1% bovine gamma
(10) Liu, Y.; Gray, N. S. Rational design of inhibitors that bind to
inactive kinase conformations. Nat. Chem. Biol. 2006, 2, 358–364.
(11) Ma, H.; Deacon, S.; Horiuchi, K. The challenge of selecting protein
kinase assays for lead discovery optimization. Expert Opin. Drug
Discovery 2008, 3, 607–621.

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 1 367
(12) Olive, D. M. Quantitative methods for the analysis of protein
phosphorylation in drug development. Expert Rev. Proteomics
2004, 1, 327–341.
(13) Getlik, M.; Grutter, C.; Simard, J. R.; Kluter, S.; Rabiller, M.;
Rode, H. B.; Robubi, A.; Rauh, D. Hybrid compound design to
overcome the gatekeeper T338M mutation in cSrc. J. Med. Chem.
2009, 3915–3926.
(28) Blencke, S.; Zech, B.; Engkvist, O.; Greff, Z.; Orfi, L.; Horvath, Z.;
Keri, G.; Ullrich, A.; Daub, H. Characterization of a conserved
structural determinant controlling protein kinase sensitivity to
selective inhibitors. Chem. Biol. 2004, 11, 691–701.
(29) Daub, H.; Specht, K.; Ullrich, A. Strategies to overcome resistance
to targeted protein kinase inhibitors. Nat. Rev. Drug Discovery
2004, 3, 1001–1010.
€
€
(14) Naqvi, T.; Rouhani, R.; Singh, R. Improved Receptor Detection.
(30) Shaw, D.; Wang, S. M.; Villasenor, A. G.; Tsing, S.; Walter, D.;
Browner, M. F.; Barnett, J.; Kuglstatter, A. The crystal structure of
JNK2 reveals conformational flexibility in the MAP kinase insert
and indicates its involvement in the regulation of catalytic activity.
J. Mol. Biol. 2008, 383, 885–893.
(31) Bukhtiyarova, M.; Northrop, K.; Chai, X.; Casper, D.; Karpusas,
M.; Springman, E. Improved expression, purification, and crystal-
lization of p38alpha MAP kinase. Protein Expression Purif. 2004,
37, 154–161.
(32) Kabsch, W. Automatic processing of rotation diffraction data
from crystals of initially unknown symmetry and cell constants.
J. Appl. Crystallogr. 1993, 26, 795–800.
(33) Read, R. J. Pushing the boundaries of molecular replacement with
maximum likelihood. Acta Crystallogr., Sect. D: Biol. Crystallogr.
2001, 57, 1373–1382.
(34) Michelotti, E. L.; Moffett, K. K.; Nguyen, D.; Kelly, M. J.; Shetty,
R.; Chai, X.; Northrop, K.; Namboodiri, V.; Campbell, B.;
Flynn, G. A.; Fujimoto, T.; Hollinger, F. P.; Bukhtiyarova, M.;
Springman, E. B.; Karpusas, M. Two classes of p38alpha MAP
kinase inhibitors having a common diphenylether core but exhibit-
ing divergent binding modes. Bioorg. Med. Chem. Lett. 2005, 15,
5274–5279.
WO 2003102154 (A2), 2003.
(15) Zaman, G. J.;vanderLee, M. M.;Kok, J. J.;Nelissen,R. L.;Loomans,
E. E. Enzyme fragment complementation binding assay for p38alpha
mitogen-activated protein kinase to study the binding kinetics of
enzyme inhibitors. Assay Drug Dev. Technol. 2006, 4, 411–420.
(16) Naqvi, T.; Rouhani, R.; Singh, R. Short Enzyme Fragment Donor.
WO 03093786 (A2), 2006.
(17) Fabian, M. A.; Biggs, W. H., 3rd; Treiber, D. K.; Atteridge, C. E.;
Azimioara, M. D.; Benedetti, M. G.; Carter, T. A.; Ciceri, P.;
Edeen, P. T.; Floyd, M.; Ford, J. M.; Galvin, M.; Gerlach, J. L.;
Grotzfeld, R. M.; Herrgard, S.; Insko, D. E.; Insko, M. A.; Lai, A.
G.; Lelias, J. M.; Mehta, S. A.; Milanov, Z. V.; Velasco, A. M.;
Wodicka, L. M.; Patel, H. K.; Zarrinkar, P. P.; Lockhart, D. J.
A small molecule-kinase interaction map for clinical kinase
inhibitors. Nat. Biotechnol. 2005, 23, 329–336.
(18) Karaman, M. W.; Herrgard, S.; Treiber, D. K.; Gallant, P.; Atteridge,
C. E.; Campbell, B. T.; Chan, K. W.; Ciceri, P.; Davis, M. I.; Edeen,
P. T.; Faraoni, R.; Floyd, M.; Hunt, J. P.; Lockhart, D. J.; Milanov,
Z. V.; Morrison, M. J.; Pallares, G.; Patel, H. K.; Pritchard, S.;
Wodicka, L. M.; Zarrinkar, P. P. A quantitative analysis of kinase
inhibitor selectivity. Nat. Biotechnol. 2008, 26, 127–132.
(19) Pargellis, C.; Tong, L.; Churchill, L.; Cirillo, P. F.; Gilmore, T.;
Graham, A. G.; Grob, P. M.; Hickey, E. R.; Moss, N.; Pav, S.;
Regan, J. Inhibition of p38 MAP kinase by utilizing a novel
allosteric binding site. Nat. Struct. Biol. 2002, 9, 268–272.
(20) Cirillo, P. F.; Pargellis, C.; Regan, J. The non-diaryl heterocycle classes
of p38 MAP kinase inhibitors. Curr Top Med Chem 2002, 2, 1021–1035.
(21) Sullivan, J. E.; Holdgate, G. A.; Campbell, D.; Timms, D.;
Gerhardt, S.; Breed, J.; Breeze, A. L.; Bermingham, A.; Pauptit,
R. A.; Norman, R. A.; Embrey, K. J.; Read, J.; VanScyoc, W. S.;
Ward, W. H. Prevention of MKK6-dependent activation by bind-
ing to p38alpha MAP kinase. Biochemistry 2005, 44, 16475–16490.
(22) Regan, J.; Breitfelder, S.; Cirillo, P.; Gilmore, T.; Graham, A. G.;
Hickey, E.; Klaus, B.; Madwed, J.; Moriak, M.; Moss, N.; Pargellis,
C.; Pav, S.; Proto, A.; Swinamer, A.; Tong, L.; Torcellini, C. Pyrazole
urea-based inhibitors of p38 MAP kinase: from lead compound to
clinical candidate. J. Med. Chem. 2002, 45, 2994–3008.
(35) Emsley, P.; Cowtan, K. Coot: model-building tools for molecular
graphics. Acta Crystallogr. D 2004, 60, 2126–2132.
€
(36) Brunger, A. T.; Adams, P. D.; Clore, G. M.; DeLano, W. L.; Gros,
P.; Grosse-Kunstleve, R. W.; Jiang, J. S.; Kuszewski, J.; Nilges, M.;
Pannu, N. S.; Read, R. J.; Rice, L. M.; Simonson, T.; Warren, G. L.
Crystallography & NMR system: a new software suite for macro-
molecular structure determination. Acta Crystallogr., Sect. D: Biol.
Crystallogr. 1998, 54, 905–921.
(37) Murshudov, G. N.; Vagin, A. A.; Dodson, E. J. Refinement of
macromolecular structures by the maximum-likelihood method.
Acta Crystallogr. D 1997, 53, 240–255.
€
(38) Schuttelkopf, A. W.; van Aalten, D. M. PRODRG: a tool for high-
throughput crystallography of protein-ligand complexes. Acta
Crystallogr., Sect. D: Biol. Crystallogr. 2004, 60, 1355–1363.
(39) Laskowski, R. A.; McArthur, M. W.; Moss, D. S.; Thornton, J. M.
PROCHECK: a program to check the stereochemical quality of
protein structures. J. Appl. Crystallogr. 1993, 26, 263–291.
(40) DeLano, W. L. The PyMOL Molecular Graphics System. http://
www.pymol.org, 2002.
(41) Dominguez, C.; Powers, D. A.; Tamayo, N. p38 MAP kinase
inhibitors: many are made, but few are chosen. Curr. Opin. Drug
Discovery Dev. 2005, 8, 421–430.
(42) Lee, M. R.; Dominguez, C. MAP kinase p38 inhibitors: clinical
results and an intimate look at their interactions with p38alpha
protein. Curr. Med. Chem. 2005, 12, 2979–2994.
(43) Hanke, J. H.; Gardner, J. P.; Dow, R. L.; Changelian, P. S.;
Brissette, W. H.; Weringer, E. J.; Pollok, B. A.; Connelly, P. A.
Discovery of a novel, potent, and Src family-selective tyrosine
kinase inhibitor. Study of Lck- and FynT-dependent T cell activa-
tion. J. Biol. Chem. 1996, 271, 695–701.
€
(23) Simard, J. R.; Getlik, M.; Grutter, C.; Pawar, V.; Wulfert, S.;
Rabiller, M.; Rauh, D. Development of a fluorescent-tagged kinase
assay system for the detection and characterization of allosteric
kinase inhibitors. J. Am. Chem. Soc. 2009, 131, 13286–13296.
€
€
(24) Simard, J. R.; Kluter, S.; Grutter, C.; Getlik, M.; Rabiller, M.;
Rode, H. B.; Rauh, D. A new screening assay for allosteric
inhibitors of cSrc. Nat. Chem. Biol. 2009, 5, 394–396.
(25) Frantz, B.; Klatt, T.; Pang, M.; Parsons, J.; Rolando, A.; Williams,
H.; Tocci, M. J.; O’Keefe, S. J.; O’Neill, E. A. The activation state
of p38 mitogen-activated protein kinase determines the efficiency
of ATP competition for pyridinylimidazole inhibitor binding.
Biochemistry 1998, 37, 13846–13853.
(26) Seeliger, M. A.; Nagar, B.; Frank, F.; Cao, X.; Henderson, M. N.;
Kuriyan, J. c-Src binds to the cancer drug imatinib with an inactive
Abl/c-Kit conformation and a distributed thermodynamic penalty.
Structure 2007, 15, 299–311.
(44) Juers, D. H.; Jacobson, R. H.; Wigley, D.; Zhang, X. J.; Huber, R.
E.; Tronrud, D. E.; Matthews, B. W. High resolution refinement of
beta-galactosidase in a new crystal form reveals multiple metal-
binding sites and provides a structural basis for alpha-complemen-
tation. Protein Sci. 2000, 9, 1685–1699.
(27) Zhang, J. H.; Chung, T. D.; Oldenburg, K. R. A simple statistical
parameter for use in evaluation and validation of high throughput
screening assays. J Biomol. Screening 1999, 4, 67–73.