Affinity Reagents that Target a Specific Inactive Form of Protein Kinases

Article abstract of DOI:10.1016/j.chembiol.2010.01.008

A number of small-molecule inhibitors have been developed that target the catalytic domains of protein kinases that are not in an active conformation. An inactive form that has been observed in several kinases is the DFG-out conformation. This conformation is characterized by an almost 180° rotation of the conserved Asp-Phe-Gly (DFG) motif in the ATP-binding cleft relative to the active form. However, the sequence and structural determinants that allow a kinase to stably adopt the DFG-out conformation are not known. Here, we characterize a series of inhibitors based on a general pharmacophore for this inactive form. We demonstrate that modified versions of these inhibitors can be used to study the thermodynamics and kinetics of ligand binding to DFG-out-adopting kinases and for enriching these kinases from complex protein mixtures.

Full text of DOI:10.1016/j.chembiol.2010.01.008

Chemistry & Biology
Article
Affinity Reagents that Target a Specific
Inactive Form of Protein Kinases
Pratistha Ranjitkar,1 Amanda M. Brock,1 and Dustin J. Maly1,
*
1Department of Chemistry, University of Washington, Seattle, WA 98195, USA
*Correspondence: maly@chem.washington.edu
DOI 10.1016/j.chembiol.2010.01.008
SUMMARY
phosphate transfer and an inactive state (or states) with reduced
catalytic activity (Huse and Kuriyan, 2002). The conformation of
A number of small-molecule inhibitors have been the ATP-binding site in these two states can be quite distinct,
characterized by large conformational shifts in several con-
served catalytic residues. A number of phosphorylation events
developed that target the catalytic domains of
protein kinases that are not in an active conforma-
tion. An inactive form that has been observed in
several kinases is the DFG-out conformation. This
and protein-protein interactions regulate the conformational
state of these enzymes (Jeffrey et al., 1995; Pellicena and Kur-
iyan, 2006; Wong et al., 2004).
conformation is characterized by an almost 180ꢀ
The ATP-binding sites of protein kinases in the active confor-
mation are very similar, while the structures of inactive kinases
rotation of the conserved Asp-Phe-Gly (DFG) motif
in the ATP-binding cleft relative to the active form.
However, the sequence and structural determinants
are much more varied (Engh and Bossemeyer, 2001; Kornev
et al., 2006). A specific inactive form of the ATP-binding cleft
that allow a kinase to stably adopt the DFG-out
conformation are not known. Here, we characterize
that has been observed in a number of kinases is the DFG-out
conformation (Liu and Gray, 2006; Okram et al., 2006). Several
a series of inhibitors based on a general pharmaco- potent kinase inhibitors have been found to selectively bind to
the DFG-out conformation of their kinase targets (type II inhibi-
tors). In many cases, these type II inhibitors exhibit a high degree
of selectivity. Despite the selectivity demonstrated by some type
phore for this inactive form. We demonstrate that
modified versions of these inhibitors can be used to
study the thermodynamics and kinetics of ligand
binding to DFG-out-adopting kinases and for enrich-
II inhibitors, a set of conserved binding interactions in the ATP-
binding pocket has been observed (Jacobs et al., 2008; Liu
ing these kinases from complex protein mixtures.
and Gray, 2006; Nagar et al., 2002; Okram et al., 2006; Simard
et al., 2009a). While the DFG-out conformation appears to be
energetically accessible to a number of kinases, the sequence
INTRODUCTION
and structural determinants that allow this transition are still
poorly understood. Furthermore, it is still not possible to predict
Comprised of over 500 members, protein kinases are the largest
which members of the kinome can be effectively targeted with
enzyme family encoded by the human genome (Manning et al.,
type II inhibitors. Identifying the full complement of kinases that
2002). These enzymes are the major mediators of intracellular
are able to stably adopt this inactive conformation will shed light
protein phosphorylation cascades and are involved in complex
on the catalytic regulation of protein kinases and may yield new
targets for drug discovery.
phenotypic responses such as growth, differentiation, adhesion,
and motility (Cohen, 2000). The proper spatial and temporal
Here we report the development of a set of affinity probes that
control of intracellular phosphorylation events by protein kinases
are derived from a general type II ligand for the DFG-out confor-
is essential for normal cellular function. Indeed, aberrant kinase
mation of protein kinases. We show that fluorophore-conjugated
activity has been implicated in numerous diseases including
analogs of these ligands are useful probes for measuring the
cancer, diabetes, and chronic inflammation (Cohen, 2002). For
thermodynamics and kinetics of type II inhibitor binding to
this reason, there has been considerable effort toward the
protein kinases. In addition, we demonstrate that these reagents
development of pharmacological agents that target this
are capable of selectively enriching protein kinases that can
enzyme family.
adopt this inactive form from complex protein mixtures. Using
The overall structure of the catalytic domain of protein kinases
these probes we provide evidence that the STE20 kinase LOK
is highly conserved (Figure 1A) (Hanks and Hunter, 1995). It
behaves like a kinase that can adopt the DFG-out conformation.
consists of 250–300 residues that form a bi-lobal structure con-
taining a smaller, mostly b stranded, N-terminal domain and
a larger, mostly a-helical, C-terminal domain (Taylor et al., RESULTS
1988). Adenosine 50-triphosphate (ATP) sits in a narrow hydro-
phobic cleft between the N- and C-terminal lobes, with the A General Type II Inhibitor Scaffold
purine moiety forming a pair of hydrogen bonds with the hinge In order to develop a set of reagents for investigating inactive
region (Engh and Bossemeyer, 2001). Most protein kinases forms of protein kinases we wished to identify a general ligand
possess at least two distinct catalytic states: an active state for the DFG-out conformation. Recently, we have characterized
where all of the catalytic residues are in an optimal position for a series of triazolopyridines that are equipotent inhibitors of the
Chemistry & Biology 17, 195–206, February 26, 2010 ª2010 Elsevier Ltd All rights reserved 195

Chemistry & Biology
Affinity Reagents for Inactive Protein Kinases
B
Gatekeeper
residue
A
Glu 310
HN
N-Lobe
O
helix-αC
Thr 338
Hinge Region
R
NH
HO
Adenine
Pocket
DFG-out Pocket
DFG-out
Hydrophobic
HN
O
O
Pocket II
R
R
N
DFG-out Pocket
H
D
HN
O
B
E
C
H
N
Adenine Pocket
O
A
C-Lobe
O
HN
O
Asp 404
Hydrophobic
Pocket I
Phe 405
DFG-motif
C
D
IC50 (nM)
< 25
1a 1b 1c 2a 2d 3a
25-100
100 -1000
>1000
R =
ABL
SRC
HCK
O
a
b
CF3
HN
HN
N
H
= Not Tested
N
N
R
HN
O
N
N
O
O
O
N
N
N
H
LCK
p38
N
N
1
c
HN
N
H
CSK
CF3
c-KIT
PYK2
PKA
N
N
R
O
a
d
HN
N
N
CF3
HN
HN
N
H
O
O
O
CLK1
PAK4
PAK5
2
N
N
R
HN
O
N
O
CF3
a
HN
N
H
MAP3K5
IRAK4
3
ERK2
196 Chemistry & Biology 17, 195–206, February 26, 2010 ª2010 Elsevier Ltd All rights reserved

Chemistry & Biology
Affinity Reagents for Inactive Protein Kinases
tyrosine kinases SRC and ABL (Seeliger et al., 2009). Through 2007). The results of the inhibition assays are shown in
structural and kinetic studies we demonstrated that this class Figure 1D. SRC, LCK, ABL, PYK2, CSK, HCK, p38, and c-KIT
of inhibitors binds to the DFG-out conformation of these kinases. were potently inhibited by most of the type II inhibitors (1a–1c,
The general binding mode of the triazolopyridine ligands is similar 2a, and 3a), while kinases that have not previously been
to other type II inhibitors that have been described (Figure 1B; observed in the DFG-out conformation show little or no inhibition
PDB IDs: 3G6G and 3G6H). The aniline and triazine groups (rings at the highest concentration tested (1 mM). Consistent with these
A and B) form a pair of hydrogen bonds with the hinge region and inhibitors binding to the DFG-out conformation, control com-
occupy the hydrophobic cleft used by adenine. Rings C and D pound 2d shows minimal inhibition of kinases that are sensitive
serve as a linker connecting the adenine-binding region to to type II inhibitors. Surprisingly, the identity of the aryl group
a hydrophobic pocket (DFG-out pocket) that is created by the that projects into the DFG-out pocket has only a small impact
nearly 180ꢀ rotation of the Phe in the DFG motif to a more on the selectivity profile among DFG-out-adopting kinases.
solvent-exposed position. Upon entering the DFG-out pocket,
the amide group connecting rings D and E forms a pair of BODIPY-Labeled Compounds as Binding Probes
hydrogen bonds with a conserved glutamate in the helix aC for Protein Kinases
and the backbone of the DFG motif. Finally, ring E occupies the In most cases, the potencies of compounds that bind the DFG-
DFG-out pocket. A small panel of inhibitors based on the triazo- out inactive conformation have been determined by their ability
lopyridine scaffold was generated to determine the generality of to inhibit activated kinases. However, this method of analysis
this ligand class (Figure 1C). For inhibitor series 1 and 2, the does not allow kinase forms with low catalytic activity to be
substituents that occupy the DFG-out pocket and exit the probed and it is often difficult to correlate the observed inhibition
adenine-binding site were varied (rings A and E), while the core to the binding of a specific kinase conformation. A direct binding
scaffold (rings B, C, and D) was held constant. Compounds 1a assay has the ability to measure ligand affinity for any activa-
and 2a contain a 3-trifluoromethylbenzamide group, which is tion state of a protein kinase and allows binding kinetics to be
a common binding element in many type II inhibitors (Liu and determined. For this reason, we determined the feasibility of con-
Gray, 2006; Okram et al., 2006). Compounds 1b and 1c contain verting our general type II inhibitors into fluorescently-labeled
functional groups that are components of selective type II inhib- binding probes. Based on a previously reported crystal structure
itors: 1b contains a 2-morphilinopyridyl group, which is the phar- of inhibitor 2a bound to SRC, the solvent-exposed 4-position of
macophore that occupies the DFG-out pocket in the selective ring A was selected as the site of fluorophore attachment
type II p38 MAPK inhibitor MPAQ, and compound 1c contains (Figure 2A). Because inhibitors 1a and 2a have similar potencies
the functional groups that are present in the selective type II against most kinases, a BODIPY-FL fluorophore was conjugated
BCR-ABL inhibitor nilotinib (Cumming et al., 2004; Sullivan to the monoalkoxy derivative 1a. To determine whether fluoro-
et al., 2005; Weisberg et al., 2005). Finally, compound 2d was phore attachment significantly affects the potency of this inhib-
generated as a control compound due to its lack of a substituent itor against protein kinases, in vitro activity assays were per-
that occupies the DFG-out pocket. In addition to the triazolopyr- formed with p38 and SRC. Gratifyingly, the IC50 of probe 4a for
idine inhibitors from series 1, a pyrimidinepyridine compound (3a) SRC and p38 is similar to inhibitor 1a (Figure 2B).
was also generated to determine the role of scaffold structure.
We hypothesized that probe 4a would show an increase in
To determine whether the inhibitors generated are potent fluorescence when bound to the ATP-binding cleft of a protein
ligands for kinases that adopt the DFG-out conformation, they kinase. To test this, 10 nM of probe 4a was incubated with
were tested against a panel of kinases in which high resolution increasing concentrations of the catalytic domains of ABL or
crystal structures of the catalytic domain are available. Crystal p38 (1–250 nM). In the presence of these kinases, probe 4a
structures of the tyrosine kinases SRC, LCK, ABL, PYK2, CSK, demonstrated a concentration-dependent increase in fluores-
and c-KIT and the serine/threonine kinase p38a in the DFG-out cence (Figure 2C, top). To confirm that the observed increase
conformation have been obtained and these kinases have in fluorescence is due to a specific interaction in the ATP-binding
been found to be sensitive to type II inhibitors (Han et al., 2009; pocket, this experiment was repeated in the presence of an
Mol et al., 2004; Seeliger et al., 2009; Simard et al., 2009b). In excess of a competitor that does not contain a fluorophore
contrast, the serine/threonine kinases PKA, PAK4, PAK5, (1 mM of 1a). As expected, no increase in fluorescence was
CLK1, MAP3K5, IRAK4, and ERK2 have not been observed in observed for either kinase in the presence of excess competitor
this inactive form and inhibitor selectivity screens have shown (Figure 2C). Having validated our binding assay with p38 and
them to only be sensitive to type I inhibitors (Fedorov et al., ABL, the Kd values of probe 4a for a panel of kinases was
Figure 1. The DFG-Out Conformation of Protein Kinases and Type II Inhibitors
(A) The catalytic domain of the tyrosine kinase SRC bound to the type II inhibitor 2a. The ATP-binding cleft is in the DFG-out conformation. The conserved gluta-
mate in helix aC (shown in yellow) and the conserved aspartate and phenylalanine in the DFG motif (shown in black) are in conformations observed for kinases
bound to type II inhibitors. Kinase subpockets occupied by type II inhibitors are indicated by dashed circles.
(B) A schematic representation of 2a bound to the tyrosine kinase SRC. This inhibitor forms two hydrogen bonds with the hinge region of the kinase. The char-
acteristic set of hydrogen bonds that type II inhibitors form with the conserved glutamate residue in the helix aC and the amide backbone of the aspartate of the
DFG motif are shown. The DFG-out pocket that is generated by the movement of the phenylalanine residue of the DFG motif is shaded.
(C) Chemical structures of inhibitors 1a-1c, 2a, 2d, and 3a.
(D) IC50s of the inhibitors shown in (C) against a diverse panel of kinases. Potencies are represented in grayscale. Note: We have previously reported the IC50s of
inhibitors 1a and 2a for SRC, ABL, and HCK (Seeliger et al., 2009). In this publication these compounds are referred to as DSA7 and DSA1, respectively. All assays
were run in triplicate or quadruplicate.
Chemistry & Biology 17, 195–206, February 26, 2010 ª2010 Elsevier Ltd All rights reserved 197

Chemistry & Biology
Affinity Reagents for Inactive Protein Kinases
A
B
O
N
N
HN
N
H
HN
O
N
N
CF3
1a
O
N
N
HN
N
H
HN
O
N
N
CF3
IC50 (nM)
25.4 3.1
SRC
p38 25.5 4.8
NH
H3C
O
F
B
O
4
F
N
N
N
CH3
H
4a
C
D
6000
4000
2000
0
ABL
ABL + 1 μM 1a
p38
p38 + 1 μM 1a
20000
15000
10000
5000
0
ABL
SRC
p38
-2000
0
50
100
[Kinase] nM
150
200
250
6000
4000
2000
0
SRC
CSK
LCK
PAK4
CLK1
MAP3K5
0
200
400
Time (min)
600
800
-2000
0
200
400
600
800
1000
[Kinase] nM
Figure 2. Fluorophore-Labeled Ligands for Studying the Kinetics and Thermodynamics of Type II Inhibitor Binding
(A) Crystal structure of 2a bound to the catalytic domain of SRC T338I (PDB ID: 3G6H). DFG motif, magenta; helix aC, teal.
(B) Conversion of compound 1a into a BODIPY-labeled binding probe (4a). IC50 values of 4a against SRC and p38 in activity assays.
(C) (Top) Change in fluorescence observed with increasing concentrations of kinases (ABL, green; p38, purple) in the presence of fluorescent probe 4a (10 nM)
and in the presence of both 4a (10 nM) and competitor 1a (1000 nM). (Bottom) Change in fluorescence observed with increasing concentrations of the kinases
SRC, CSK, LCK, PAK4, CLK1, and MAP3K5 in the presence of 4a (10 nM).
198 Chemistry & Biology 17, 195–206, February 26, 2010 ª2010 Elsevier Ltd All rights reserved

Chemistry & Biology
Affinity Reagents for Inactive Protein Kinases
Table 1. Kd Values of Kinases Determined with 5 nM (ABL, SRC,
and HCK) or 10 nM (LCK, p38, CSK, EPHA3, CLK1, PAK4, PAK5,
and MAP3K5) 4a
Table 2. Kd, t1/2, and koff Values for Imatinib-Resistant Mutants of
ABL and the Corresponding SRC Mutants
Mutants
Kd (nM)
t1/2 (min)
koff (sꢁ1
)
5.2 3 10ꢁ5
4.9 3 10ꢁ5
5.9 3 10ꢁ5
7.3 3 10ꢁ5
9.0 3 10ꢁ5
6.9 3 10ꢁ5
1.8 3 10ꢁ4
1.6 3 10ꢁ4
Kd (nM)
ABL wt
9.9 1.6
220
234
195
162
3
5
7
4
ABL
9.9 1.6
13
8.7 0.6
20
35 11
ABL E255V
ABL Y253H
SRC wt
45
4
SRC
2
9.4 1.2
HCK
13
15
12
11
2
2
3
1
LCK
2
SRC E280V
SRC F278H
SRC T338I
p38
133 18
p38
167
67
4
9
1
CSK
17
27
3
5
EPHA3
35 11
72
CLK1
>1000
>1000
>1000
>1000
Kd values for ABL and SRC were determined with 10 nM 4a. Values
shown are the average of three assays SEM.
PAK4
PAK5
MAP3K5
Cardama et al., 2007). Consistent with inhibitors of this class not
making significant interactions with this flexible region of the
kinase, the Y253H mutant of ABL was found to have an identical
binding affinity for 4a as the wild-type enzyme. However, the
E255V mutant demonstrated a 4- to 5-fold weaker interaction.
The analogous mutants of SRC, F278H, and E280V were also
characterized. The less flexible nature of SRC’s P-loop is re-
flected in the identical kinetics and thermodynamics observed
for both mutants and wild-type. Finally, a gatekeeper mutant of
SRC, SRC T338I, was characterized. Similar to the underivatized
inhibitor, 4a has an equal affinity for wild-type and SRC T338I.
Despite this similarity in binding thermodynamics, the on
and off rates for 4a are two to three times faster for the gate-
keeper mutant.
Values shown are the average of three assays SEM.
determined (Figure 2C, bottom). Consistent with the in vitro
activity assays described above, kinases that have previously
been demonstrated to be able to adopt the DFG-out conforma-
tion (ABL, SRC, HCK, LCK, p38, EPHA3, and CSK) have a high
affinity for probe 4a, while other kinases (CLK1, PAK4, PAK5,
and MAP3K5) do not show a detectable level of binding (Table 1).
Because a large conformational change in the ATP-binding
pocket of kinases is required to accommodate inhibitor binding,
most type II inhibitors demonstrate slow binding kinetics. To
determine if 4a can be used to obtain type II inhibitor binding
kinetics, an assay that allows the measurement of the rate of
kinase-4a complex dissociation was developed. For this assay, Generation of Affinity Matrices 14a and 15a
the catalytic domains of SRC, ABL, and p38 were incubated with We determined the feasibility of converting compounds 1a and
probe 4a under conditions in which >95% of the fluorescent 3a into affinity pull-down reagents. These proteomic tools should
probe was bound. The reaction mixture was then diluted allow unbiased analyses of the kinases that are able to adopt the
30-fold into a solution containing a large excess of competitor DFG-out conformation in diverse cellular lysates. The 4-position
1a, to ensure that dissociated 4a cannot re-bind to the kinase. of ring A was modified with a flexible polyethylene glycol linker
By measuring the time-dependent loss in fluorescence, the that contains a primary amino group to allow selective immobili-
rate of dissociation of the SRC-4a, ABL-4a, and p38-4a zation to a solid support (compounds 12a and 13a; Figure 3A).
complexes were determined (Figure 2D). The data obtained Triazolopyridine 12d, which is an analog of control compound
under these conditions fit well to a monoexponential decay 2d, was also generated as a negative control for enrichment
curve, with t1/2 for the SRC, ABL, and p38 complexes of 162, studies. To determine whether linker attachment affects the
220, and 72 min, respectively (Table 2). These data correlate potency of this scaffold for kinases that adopt the DFG-out
well with kinetic data that has been obtained for this class of conformation, in vitro activity assays were performed (Figure 3C).
compounds and other type II inhibitors (Pargellis et al., 2002; Gratifyingly, attachment of a flexible polyethylene glycol linker
Seeliger et al., 2007, 2009; Sullivan et al., 2005).
does not alter the selectivity or potency of these inhibitors. Inhib-
We next investigated how mutations in the catalytic domains itor-tethered resins containing 12a and 13a were generated by
of protein kinases affect the conformational dynamics of the the procedure shown in Figure 3B. In each coupling reaction,
ATP-binding cleft. Specifically, probe 4a was used to study the corresponding non-linkable analogs (1a, 3a, or 2d) were
how imatinib-resistance mutations in the P-loop affect the inter- included as controls. Addition of this internal standard allowed
action of ABL kinase with type II inhibitors (Table 2). The Kd of this the progress of the coupling reaction to be monitored by
probe for two resistance mutants that are frequently observed in observing the selective loss of the amine-containing compounds
patients with chronic myelogenous leukemia that are undergoing 12a, 13a, and 12d by HPLC (see Figures S1 and S2 available
imatinib treatment, Y253H and E255V, was determined (Quintas- online).
(D) Determination of the SRC-4a, ABL-4a, and p38-4a dissociative half-lives (t1/2). The catalytic domains of these kinases were incubated under conditions in
which >95% of 4a (500 nM) is bound. The complexes were then diluted 30-fold into a solution containing an excess of competitor 1a (5 mM) to prevent reformation
of the kinase-probe complex. The rate of dissociation is measured by the decrease in fluorescence over time. Fluorescence was read every 10 min for 8 hr, and
then at 12 hr. Data was fit to a monoexponential decay curve. Assays were run in triplicate.
Chemistry & Biology 17, 195–206, February 26, 2010 ª2010 Elsevier Ltd All rights reserved 199

Chemistry & Biology
Affinity Reagents for Inactive Protein Kinases
A
N
N
N
Cl
Cl
O
O
Cl
Cl
N
N
N
N
N
HN
N
H
N
N
HN
N
H
5
a-b
c-e
HN
x
N
HN
O
x
N
CF3
or
N
CF3
X = N, 12a
X = N, 10a
X = CH, 13a
X = CH, 11a
O
O
6
HN
Boc
N
H
NH2
O
O
3
Reagents and Conditions: (a) if X = N; (i) 7, i-PrOH, rt; (ii) TEA, rt; if X= CH; 7, EtOH, 80°C; (b) 8, Et3N-TFA, DMSO, 95°C; (c) 30%
TFA, CH2Cl2, rt; (d) 9, HOBt, EDC, DIEA, CH2Cl2, rt; (e) 30% TFA, CH2Cl2, rt
NH2
O
Boc
O
O
OH
H2N
N
H
N
H
O
O
O
O
9
8
Boc
CF3
N
H
7
N
N
HN
B
HN
N
N
12a, 13a or 12d
OH
O
OH
O
H
H
N
EDC, DIEA
N
R1
50:50 DMF/EtOH
OH
N
5
5
H
12d
O
O
R1 = 12a, 13a, 12d
HN
NH2
O
O
3
C
IC (nM)
50
ABL
SRC
HCK
LCK p38
CSK
PKA
CLK1
PAK4
PAK5 MAP3K5 IRAK4
ERK2
12a 6.6 0.5 3.3 0.6 9.6 0.4 < 10 < 13 16.3 1.6 > 10000 > 10000 > 10000 > 10000 > 10000 > 10000 > 10000
13a 6.6 1.0 5.7 3.1 6.3 1.4 < 10 < 13 140 > 10000 > 10000 > 10000 > 10000 > 10000 > 10000 > 10000
9
Figure 3. Synthesis and Characterization of Immobilized type II Inhibitors
(A) Synthetic scheme for generating type II inhibitors that can be immobilized.
(B) Procedure for generating affinity matrices 14a (R1 = 12a), 14d (R1 = 12d), and 15a (R1 = 13a).
(C) IC50s of inhibitors 12a and 13a against a panel of kinases.
Affinity Matrices 14a and 15a Enrich Kinases that Adopt
the DFG-Out Conformation
4A and 4B). Because E. coli lack eukaryotic and eukaryotic-like
protein kinases, only exogenously added kinases should be re-
A model system was developed to determine the ability of affinity tained by the affinity matrix (Kannan et al., 2007). We first tested
matrices 14a and 15a to enrich kinases that are able to adopt the the ability of affinity matrix 15a to enrich p38 kinase from lysates.
DFG-out conformation. Optimal conditions for affinity capture, A majority of His6-p38 was retained by affinity matrix 15a after
resin washing, and protein elution were determined by adding 2 hr of incubation as demonstrated by the lack of detectable
purified protein kinases to cellular lysates from E. coli (Figures levels of His6-tagged kinase in the soluble fraction (Figure 4A).
200 Chemistry & Biology 17, 195–206, February 26, 2010 ª2010 Elsevier Ltd All rights reserved

Chemistry & Biology
Affinity Reagents for Inactive Protein Kinases
A
B
p38
p38
MAP3K5
PAK5
Probe 15a
Probe 15a
Probe 15a
Probe 14d
Probe 14d
Probe 14a
Probe 14a
ABL
SRC
HCK
ABL
HCK
50 kDa
35 kDa
35 kDa
15 kDa
p38
SRC
50 kDa
35 kDa
50 kDa
35 kDa
50 kDa
35 kDa
p38
ABL
HCK
C
HEK-293 Cell Lysate
Figure 4. Characterization of Immobilized Type II Inhibitors 14a and 15a
(A) Enrichment of protein kinases from E. coli lysates with affinity resins 14a or 15a or control matrix 14d. Purified His6-p38 (4.6 mg), His6-ABL (5 mg), or HCK
(18 mg) were added to E. coli lysate (550–700 mg) and subjected to standard enrichment conditions with affinity matrix 15a or control matrix 14d. (Top) All fractions
were subjected to SDS-PAGE. ABL and p38 were detected with HisProbe-HRP (Pierce); SRC and HCK were detected with an antibody that specifically recog-
nizes SRC-family kinases that are dephosphorylated at Tyr416 in the activation loop [a-non-phospho-SRC(Y416)] (Cell Signaling). (Bottom) Elutions 1 and 2 from
the enrichment experiments described above were subjected to SDS-PAGE and silver stained.
(B) Enrichment of kinases that are sensitive to type II inhibitors in the presence of kinases that have not been characterized in the DFG-out conformation. His6-p38
(4 mg), His6-MAP3K5 (10 mg), and His6-PAK5 (10 mg) were added to E. coli lysate (500 mg) and subjected to standard enrichment conditions with affinity matrix
14a. The catalytic domain of SRC (4 mg), His6-MAP3K5 (4.5 mg), and His6-STK16 (4.5 mg) were added to E. coli lysate (600 mg) and subjected to standard enrich-
ment conditions with affinity matrix 14a. (Top) All fractions were subjected to SDS-PAGE and probed with HisProbe-HRP or a-non-phospho-SRC(Y416). (Bottom)
Elutions 1 and 2 from the enrichment experiments described above were subjected to SDS-PAGE and silver stained. Note: Procedures for protein kinase expres-
sion and purification are in the Supplemental Experimental Procedures.
(C) Enrichment of endogenous kinases from HEK293 cell lysate. 600 mg of HEK293 lysate was subjected to standard enrichment conditions with affinity matrix
14a. All fractions were subjected to SDS-PAGE and probed with a-non-phospho-SRC(Y416) or a-LOK.
Chemistry & Biology 17, 195–206, February 26, 2010 ª2010 Elsevier Ltd All rights reserved 201

Chemistry & Biology
Affinity Reagents for Inactive Protein Kinases
Furthermore, no immobilized p38 was found to dissociate from as other kinases that adopt the DFG-out conformation (Table
the resin after multiple washes. Repeated efforts to elute p38 S1). Furthermore, resins 14a and 15a were able to enrich the
from the affinity matrix with high concentrations of soluble, catalytic domain of LOK from E. coli lysate over other protein
ATP-competitive inhibitors were unsuccessful. This is most likely kinases (Figure 5A). In addition, LOK was not retained by control
due to the high affinity of the immobilized ligand for kinases that matrix 14d. To further confirm the affinity of LOK for the type II
adopt the DFG-out conformation and the slow rate of dissocia- inhibitor class, binding studies with probe 4a were performed.
tion of these inhibitor-kinase complexes. However, a mild dena- Similar to other kinases that adopt the DFG-out conformation,
turant (0.5% SDS) allowed quantitative elution of this kinase. LOK has a high affinity for this fluorescently-labeled probe
Furthermore, these conditions appear to be selective for disrup- (Kd = 12 nM; Figure 5B). Finally, the t1/2 for the LOK-4a complex
tion of the kinase-inhibitor complex as demonstrated by the lack was determined (Figure 5C). Consistent with LOK demonstrating
of other proteins in the eluted fractions detected via silver stain- slow binding kinetics for type II inhibitors, the t1/2 for this complex
ing. The same conditions were found to be effective for enriching was found to be 51 min. Although the off rate for probe 4a is 3- to
ABL and the SRC family kinase HCK (Figure 4A). As expected, 4-fold faster than for SRC and ABL, the kinetics observed are
none of these kinases were retained by control matrix 14d. To consistent with a conformational change being necessary for
ensure that the affinity matrix selectively enriches kinases that ligand binding.
adopt the DFG-out conformation, p38 was subjected to the stan-
dard affinity enrichment protocol in an E. coli protein lysate con- DISCUSSION
taining the kinases PAK5 and MAP3K5 (Figure 4A). As expected,
only p38 was bound and retained by affinity matrix 14a. Further- After it was discovered that the inhibitors imatinib and BIRB796
more, p38 was the only protein that could be detected in the bind to the inactive forms of their kinase targets, there has been
eluted fractions. Similar results were observed for SRC.
widespread interest in the identification of small molecules that
The ability of immobilized inhibitor 14a to enrich endogenous selectively recognize alternate active site conformations (Liu
protein kinases from mammalian lysates was tested. Western and Gray, 2006; Nagar et al., 2002; Okram et al., 2006; Schindler
blot analysis demonstrated that >90% of total SRC was retained et al., 2000). High selectivity can be achieved with inhibitors that
by the beads when 650 mg of HEK293 protein lysate was incu- bind to inactive kinase conformations and the characteristically
bated with affinity resin 14a (Figure 4C). SRC remained bound long target residence times of these protein-small molecule
to the affinity matrix during repeated washes and was eluted complexes may be beneficial from a therapeutic standpoint.
with 0.5% SDS. Comparison of the proteins eluted from control The DFG-out conformation is a specific inactive form that has
matrix 14d and affinity matrix 14a by SDS-PAGE followed by been successfully targeted by a number of type II inhibitors.
silver staining demonstrated that several unique proteins were While it was once thought that only a few kinases are able to
enriched from HEK293 and HeLa cell lysates by immobilized adopt this conformation, recent studies have demonstrated
inhibitor 14a (Figure S3). However, a number of prominent bands that it is accessible to many members of the human kinome
were present in the eluted fractions for both support-bound (Choi et al., 2009; Dar et al., 2008; DiMauro et al., 2006; Han
ligands. These appear to be proteins that bind nonspecifically et al., 2009; Hodous et al., 2007; Mol et al., 2004; Schroeder
to the matrix, as preincubation with control matrix 14d prior to et al., 2009; Seeliger et al., 2009; Wan et al., 2004). However,
enrichment with 14a did not deplete a majority of these bands. the sequence and structural determinants that allow a kinase
Mass spectrometric analysis of the proteins that were enriched to adopt the DFG-out conformation are still not known and it is
from HEK293 and HeLa cells is ongoing and will be reported not possible to predict which kinases will be sensitive to type II
elsewhere. In parallel, a candidate approach for identifying inhibitors. To address this issue, our goal was to identify a general
kinases that can adopt the DFG-out conformation was pursued. ligand for kinases that adopt the DFG-out conformation. We
Previous inhibitor screens have shown that the serine/threonine have found that compounds based on the triazolopyridine scaf-
kinase LOK is sensitive to several type II inhibitors (Angell et al., fold, which we have previously demonstrated through kinetic
2008; Fabian et al., 2005; Karaman et al., 2008). However, and structural studies to bind the DFG-flipped forms of the
biochemical studies have not been performed to establish kinases ABL and SRC, potently inhibit a broad spectrum of
whether LOK can adopt the DFG-out conformation. To deter- kinases that are sensitive to type II inhibitors. Furthermore, these
mine whether endogenous LOK binds to the triazolopyridine compounds do not inhibit other well-characterized kinases that
scaffold, HEK293 cell lysate was incubated with affinity matrix have not previously been observed in this specific inactive form.
14a. Immunoblot analysis of the bound and unbound fractions
A small panel of analogs based on the triazolopyridine scaffold
with an a-LOK antibody demonstrated that a percentage of was generated to determine how substituents that occupy the
this kinase was retained by the beads and selectively eluted.
pocket created by the DFG motif flipping to a more solvent
exposed position affect kinase selectivity (Figure 1). As ex-
pected, a control compound that lacks a hydrophobic moiety
that occupies the DFG-out pocket and an amide group that
The Catalytic Domain of LOK Behaves Similarly
to Other DFG-Out-Adopting Protein Kinases
To further characterize LOK, the His6-tagged catalytic domain of makes a set of characteristic hydrogen bonds with the backbone
this kinase was expressed and purified using a previously amide of the DFG motif and a conserved glutamic acid in the
described protocol (Fedorov et al., 2007). An in vitro activity helix aC does not inhibit any of the DFG-out-adopting kinases.
assay was developed for LOK and compounds 1a-c, 2a, and Furthermore, an analog that contains a 3-trifluormethylbenza-
3a were tested for their ability to inhibit this kinase. LOK was mide (1a), which is a general pharmacophore that occupies
found to be equally sensitive to inhibition by type II inhibitors the DFG-out pocket in many type II inhibitors, appears to be
202 Chemistry & Biology 17, 195–206, February 26, 2010 ª2010 Elsevier Ltd All rights reserved

Chemistry & Biology
Affinity Reagents for Inactive Protein Kinases
A
3000
2000
1000
0
B
Probe 15a
Probe 14d
LOK
LOK
CLK1
MAP3K5
LOK
-1000
Probe 14a
0
50
100
[LOK] nM
150
200
250
C
20000
15000
10000
5000
35 kDa
35 kDa
0
0
100
200
300
400
500
LOK
LOK
Time (min)
Figure 5. The Catalytic Domain of LOK Behaves Like a Kinase that Can Adopt the DFG-Out Conformation
(A) Enrichment of the catalytic domain of LOK from E. coli lysates with affinity resins 14a or 15a or control matrix 14d. (Top) Purified His6-LOK (4.5 mg) was added
to E. coli lysate (600 mg) and subjected to standard enrichment conditions with affinity matrix 15a. LOK was detected with HisProbe-HRP (Pierce). (Middle) Puri-
fied His6-LOK (4.5 mg) was added to E. coli lysate (600 mg) and subjected to standard enrichment conditions with control matrix 14d. LOK was detected with
HisProbe-HRP. (Bottom) Purified His6-LOK (4 mg), His6-CLK1 (10 mg), and His6-MAP3K5 (10 mg) were added to E. coli lysate (500 mg) and subjected to standard
enrichment conditions with affinity matrix 14a. All fractions were subjected to SDS-PAGE and kinases were detected with HisProbe-HRP (Pierce). (Bottom gels)
Elutions 1 and 2 from the enrichment experiments performed in the top and bottom panels above. Samples were subjected to SDS-PAGE and silver stained.
(B) Change in fluorescence observed with an increasing concentration of LOK in the presence of 5 nM of fluorescent probe 4a. Kd = 12 1 nM. Value shown is the
average of three assays SEM.
(C) Determination of the LOK-4a complex dissociative half-life (t1/2). For LOK-4a, t1/2 = 51 min and koff = 2.3 3 10ꢁ5 sꢁ1. Assays were run in triplicate.
a nonselective ligand for kinases that have previously been char- trum of kinases than was previously appreciated (Angell et al.,
acterized to adopt the DFG-out conformation. Unexpectedly, re- 2008; Choi et al., 2009; Okram et al., 2006).
placing this pharmacophore with functional groups from more
To gain a better understanding of how inhibitors based on the
selective type II inhibitors does not greatly enhance the selec- triazolopyridine scaffold interact with protein kinases, we con-
tivity profile for compounds based on the triazolopyridine scaf- verted compound 1a into a fluorescently-labeled ligand that
fold. For example, compound 1b, which contains a 2-morphilino- can be used in binding assays (Figure 2). The use of a binding
pyridyl group, is a nearly equipotent inhibitor of p38, ABL, SRC, assay allows the affinity of this probe to be determined for
HCK, and CSK; however, inhibitors that display this moiety from kinases in any activation state and facilitates the direct measure-
a 4-anilinoquinazoline scaffold are >500-fold selective for p38 ment of ligand binding kinetics. As many type II inhibitors have
over these kinases (Perera and Maly, 2008). Similarly, 1c, which long residence times and slow on rates, determination of slow
contains the functional groups present in the highly selective binding kinetics with a general probe for the DFG-out conforma-
BCR-ABL inhibitor nilotinib, is a fairly nonselective inhibitor of tion can be used to verify the ability of a kinase to adopt this
DFG-out-adopting kinases. These data highlight that the ability inactive form. In the presence of DFG-out-adopting kinases,
of some type II inhibitors to discriminate between closely related fluorescent probe 4a demonstrated a concentration-dependent
kinases is most likely due to several unique protein-ligand inter- increase in fluorescence. An assay was developed to measure
actions rather than a single selectivity filter. Indeed, we have the residence time of this probe when bound to p38, SRC, and
demonstrated that a unique interaction between the P-loop of ABL. The slow binding kinetics demonstrated by this probe are
ABL and imatinib makes a major contribution to the ability of similar to those observed for other type II inhibitors of these
this compound to inhibit this kinase over closely related SRC kinases (Pargellis et al., 2002; Seeliger et al., 2007, 2009; Sullivan
(Seeliger et al., 2009). Several recent reports that describe type et al., 2005). This assay was also used to study how drug-resis-
II inhibitors with limited selectivity profiles highlight that ligands tance mutations affect the interactions of kinases with type II
of this class have the potential to inhibit a much broader spec- inhibitors. A mutation to the flexible P-loop region of ABL,
Chemistry & Biology 17, 195–206, February 26, 2010 ª2010 Elsevier Ltd All rights reserved 203

Chemistry & Biology
Affinity Reagents for Inactive Protein Kinases
E255V, was found to reduce the affinity of this kinase for the type II inhibitors that are extremely potent against this mutant
probe. Interestingly, the observed loss in affinity is due to have been described (Huang et al., 2009; O’Hare et al., 2009;
a slower on rate rather than a slower rate of dissociation. None Seeliger et al., 2009). Expanding the number and diversity of
of the analogous P-loop mutations in SRC were found to affect kinases that are known to stably adopt the DFG-out conforma-
the interaction of this kinase with type II inhibitors, which is tion will allow the identification of the sequence determinants
consistent with this region not interacting with inhibitors. Finally, that allow this transition.
mutation of the gatekeeper residue of SRC to a bulkier residue
was found to affect both the on and off rates of ligand binding.
SIGNIFICANCE
This result is consistent with recent findings that mutation of
the gatekeeper residue in ABL causes increased conformational
A number of type II inhibitors have been developed that
disturbances in a specific catalytic domain region of this kinase
(Iacob et al., 2009).
target an inactive form of the ATP-binding site of kinases
called the DFG-out conformation. Once thought to be ener-
While structural studies and kinetic assays can be used to
determine if a kinase can adopt the DFG-out conformation, these
methods can only be performed on purified proteins. Because it
getically accessible to only a few kinases, the DFG-out
conformation has been characterized with an increasing
number of enzymes from this family. However, it is still not
is difficult to know whether the conformational preferences of
a recombinantly expressed purified kinase are representative
possible to predict which members of the kinome are able
to stably adopt this inactive conformation. In addition, the
of those in a cell, probes that allow interrogation of endogenous
kinases are highly desirable. For this reason, we developed
structural and sequence determinants of this transition
remain poorly understood. Here, we have identified a scaf-
a support-bound affinity matrix based on inhibitor 1a that allows
an unbiased sampling of mammalian lysates. We first verified
that our affinity matrix was capable of enriching purified kinases
fold that appears to be a general pharmacophore for the
DFG-out conformation. Derivatives of this scaffold are
potent inhibitors of kinases that have previously been char-
that are known to be sensitive to type II inhibitors from complex
mixtures (Figure 4). Indeed, affinity matrices 14a and 15a selec-
tively enriched SRC, HCK, and p38 from an E. coli lysate but not
acterized to be sensitive to type II inhibitors. Furthermore,
fluorophore-conjugated versions of these molecules allow
the determination of the thermodynamics and kinetics of
kinases that are insensitive to type II inhibitors. Importantly,
control matrix 14d, which does not contain a pharmacophore
that occupies the DFG-out pocket, does not enrich any of the
ligand binding to any activation state of the kinase catalytic
domain. In addition, immobilized analogs of the general
scaffold were found to be effective reagents for enriching
kinases tested. We next tested this reagent in mammalian cell
lysates. First, it was verified that affinity matrix 14a was able
to enrich endogenous SRC from HEK293 cells. Analysis of the
DFG-out-adopting kinases from complex protein mixtures.
Affinity reagents 14a and 15a were used to demonstrate
that the STE20 kinase LOK behaves like other DFG-out-
proteins that were selectively enriched with affinity resin 14a over
adopting kinases. LOK is distantly related to other members
control matrix 14d from HeLa and HEK293 lysates demonstrated
that there are several unique proteins. Mass spectrometric
identification of these proteins is currently being performed.
of the kinase family that have been characterized in the DFG-
out conformation. Use of these reagents to discover the full
complement of kinases that are able to stably adopt the
A candidate approach using western blot analysis was adop-
ted to discover novel DFG-out-adopting kinases. We found that
the STE20 kinase LOK could be selectively enriched from
DFG-out conformation will allow a greater understanding
of the thermodynamics and kinetics of the catalytic domains
of this enzyme family.
HEK293 lysates with affinity matrix 14a. Furthermore, the puri-
fied catalytic domain could be enriched from E. coli lysates
EXPERIMENTAL PROCEDURES
and was sensitive (IC50 < 20 nM) to type II inhibitors 1a-1c, 2a,
and 3a in activity assays (Table S1). Finally, LOK demonstrated
similarly slow binding kinetics with type II inhibitors such as
ABL, SRC, and p38. Besides p38 and RAF, no other serine/
threonine kinases have been characterized in the DFG-out con-
formation. Furthermore, no members of the STE20 kinase family
have been found to be sensitive to type II inhibitors. Indeed,
MAP3K5, PAK4, and PAK5, which are STE20 kinases and
much more closely related to LOK than SRC or ABL, were not
sensitive to any of the affinity reagents that we have developed.
LOK is also of interest because it contains a residue with a bulky
side chain (isoleucine) at the gatekeeper position. To date, a vast
majority of the kinases that have been characterized to adopt the
DFG-out conformation (except Tie-2, PYK2, and c-MET) contain
either a threonine or valine at this position. However, the pres-
ence of a larger gatekeeper residue does not preclude a kinase
from adopting this form. While conversion of the threonione
gatekeeper of ABL to an isoleucine results in this kinase being
resistant to the type II inhibitors imatinib and nilotinib, several
Synthetic Methods
Detailed synthetic procedures are in the Supplemental Experimental Proce-
dures.
Affinity Measurements with Probe 4a
Purified kinase (initial concentration = 250 nM, 2-fold serial dilution down to
0.2 nM) was incubated with 5 or 10 nM 4a in 120 ml of buffer containing
20 mM Tris (pH 7.5), 2% DMSO, 100 mM KCl, 10% glycerol, 1 mM DTT,
0.05% Pluronic F-68, and 1 mM MgCl2. Fluorescence was read after 30 min
of incubation at room temperature (excitation = 485 nm; emission = 535 nm).
The Kd was determined by fitting the data to non-linear regression analysis
(one site total binding) with Prism GraphPad software.
Determination of t1/2 with Probe 4a
Purified kinase (700 nM) was incubated with 4a (500 nM) in 120 ml of buffer con-
taining 20 mM Tris (pH 7.5), 2% DMSO, 100 mM KCl, 10% glycerol, 1 mM DTT,
0.05% Pluronic F-68, and 1 mM MgCl2. After 4 hr, the mixture was diluted
30-fold ([4a] = 17 nM) into the same buffer containing 5 mM of 1a. Fluorescence
was read every 10 min for 8 hr, and then at 12 hr (excitation = 485 nm; emis-
sion = 535 nm). Off rates (koff) and dissociative half-lives (t1/2) were determined
204 Chemistry & Biology 17, 195–206, February 26, 2010 ª2010 Elsevier Ltd All rights reserved

Chemistry & Biology
Affinity Reagents for Inactive Protein Kinases
by fitting the data to a one phase exponential decay equation with Prism
GraphPad software.
Received: December 1, 2009
Revised: January 18, 2010
Accepted: January 20, 2010
Published: February 25, 2010
Generation of Affinity Matrix 14a
A modification of a previously published procedure was used (Wissing et al.,
2007). 1.5 ml of ECH Sepharose resin was washed with 10 bed volumes of
H2O (pH 4.5; 0.5 M NaCl) and 1:1 DMF/EtOH. 3 ml of 0.75 mM 12a and
0.46 mM 1a in 1:1 DMF/EtOH was added to the resin followed by 30 ml of
DIEA and 450 ml of 1 M EDCI in 1:1 DMF/EtOH. The reaction mixture was
rotated for 48 hr and monitored by analytical HPLC (Figures S2 and S3). After
48 hr, the reaction mixture was drained and washed with 2 bed volumes of 1:1
DMF/EtOH. The resin was then incubated with 56 ml ethanolamine, 933 ml of
20 mM acetic acid, and 450 ml of EDCI in 1:1 DMF/EtOH for 12 hr, followed
by washing with 6 bed volumes of 1:1 DMF/EtOH (33), 6 bed volumes of
0.5 M NaCl (23), and 10 bed volumes of 20% EtOH. The resin was stored at
4ꢀC in 20% EtOH until further use. Affinity matrices 14d and 15a were gener-
ated by the same method.
REFERENCES
Angell, R.M., Angell, T.D., Bamborough, P., Bamford, M.J., Chung, C.W.,
Cockerill, S.G., Flack, S.S., Jones, K.L., Laine, D.I., Longstaff, T., et al.
(2008). Biphenyl amide p38 kinase inhibitors 4: DFG-in and DFG-out binding
modes. Bioorg. Med. Chem. Lett. 18, 4433–4437.
Choi, Y., Syeda, F., Walker, J.R., Finerty, P.J., Jr., Cuerrier, D., Wojciechowski,
A., Liu, Q., Dhe-Paganon, S., and Gray, N.S. (2009). Discovery and structural
analysis of Eph receptor tyrosine kinase inhibitors. Bioorg. Med. Chem. Lett.
19, 4467–4470.
Cohen, P. (2000). The regulation of protein function by multisite phosphoryla-
tion—a 25 year update. Trends Biochem. Sci. 25, 596–601.
Cohen, P. (2002). Protein kinases—the major drug targets of the twenty-first
General Method for Affinity Enrichment from E. coli Lysate
century? Nat. Rev. Drug Discov. 1, 309–315.
50 ml of affinity matrix 14a, 14d, or 15a was washed with 8 bed volumes of
a buffer containing 50 mM HEPES (pH 7.5), 0.5% Triton X-100, 1 mM EDTA,
1 mM EGTA, and 100 mM NaCl. Purified kinase (4–18 mg) and E. coli lysate
protein (500–800 mg) in 250 ml of binding buffer [50 mM HEPES (pH 7.5),
0.5% Triton X-100, 1 mM EDTA, 1 mM EGTA, 20 mM MgCl2, 1 M NaCl, and
100 mg/ml PMSF] was added to the resin and incubated for 2 hr at 4ꢀC. The
resin was then drained, incubated with 8 bed volumes of binding buffer for
30 min at 4ꢀC, and drained again (4 washes). After the final wash, 1.5 bed
volumes of 0.5% SDS in binding buffer was added to the resin and incubated
for 10 min at room temperature. The eluted protein was collected and this
process was repeated. All collected fractions were subjected to SDS-PAGE,
transferred to nitrocellulose, and subjected to western blot analysis. Eluted
fractions were also subjected to SDS-PAGE and stained with Silver Stain
Xpress (Invitrogen).
Cumming, J.G., McKenzie, C.L., Bowden, S.G., Campbell, D., Masters, D.J.,
Breed, J., and Jewsbury, P.J. (2004). Novel, potent and selective anilinoquina-
zoline and anilinopyrimidine inhibitors of p38 MAP kinase. Bioorg. Med. Chem.
Lett. 14, 5389–5394.
Dar, A.C., Lopez, M.S., and Shokat, K.M. (2008). Small molecule recognition of
c-Src via the Imatinib-binding conformation. Chem. Biol. 15, 1015–1022.
DiMauro, E.F., Newcomb, J., Nunes, J.J., Bemis, J.E., Boucher, C., Buchanan,
J.L., Buckner, W.H., Cee, V.J., Chai, L., Deak, H.L., et al. (2006). Discovery of
aminoquinazolines as potent, orally bioavailable inhibitors of Lck: synthesis,
SAR, and in vivo anti-inflammatory activity. J. Med. Chem. 49, 5671–5686.
Engh, R.A., and Bossemeyer, D. (2001). The protein kinase activity modulation
sites: mechanisms for cellular regulation - targets for therapeutic intervention.
Adv. Enzyme Regul. 41, 121–149.
Fabian, M.A., Biggs, W.H., III, Treiber, D.K., Atteridge, C.E., Azimioara, M.D.,
Benedetti, M.G., Carter, T.A., Ciceri, P., Edeen, P.T., Floyd, M., et al. (2005).
A small molecule-kinase interaction map for clinical kinase inhibitors. Nat.
Biotechnol. 23, 329–336.
General Method for Affinity Enrichment from Mammalian Cell Lysate
50–150 ml of affinity matrix 14a, 14d, or 15a was washed with 8 bed volumes of
a buffer containing 50 mM HEPES (pH 7.5), 0.5% Triton X-100, 1 mM EDTA,
1 mM EGTA, and 100 mM NaCl. HEK293 or HeLa lysate proteins (0.6–8 mg)
in 0.25–1.0 ml of buffer containing 50 mM HEPES (pH 7.5), 0.5% Triton X-100,
1 mM EDTA, 1 mM EGTA, 2.5 mM sodium orthovanadate, 1 mM PMSF, Roche
protease inhibitor cocktail, and 20 mM MgCl2 was added to the resin and incu-
bated for 2 hr at 4ꢀC. The resin was then drained, incubated with 8 bed
volumes of binding buffer for 30 min at 4ꢀC, and drained again (4 washes). After
the final wash, 3–4 bed volumes of 0.5% SDS in binding buffer was added to
the resin and incubated for 10 min at room temperature. The eluted protein was
collected and this process was repeated. All collected fractions were lyophi-
lized, subjected to SDS-PAGE, transferred to nitrocellulose, and subjected
to western blot analysis with an a-non-phospho-SRC(Y416) or a-LOK anti-
body. Eluted fractions were subjected to SDS-PAGE and stained with Silver
Stain Xpress.
Fedorov, O., Marsden, B., Pogacic, V., Rellos, P., Muller, S., Bullock, A.N.,
Schwaller, J., Sundstrom, M., and Knapp, S. (2007). A systematic interaction
map of validated kinase inhibitors with Ser/Thr kinases. Proc. Natl. Acad.
Sci. USA 104, 20523–20528.
Han, S., Mistry, A., Chang, J.S., Cunningham, D., Griffor, M., Bonnette, P.C.,
Wang, H., Chrunyk, B.A., Aspnes, G.E., Walker, D.P., et al. (2009). Structural
characterization of proline-rich tyrosine kinase 2 (PYK2) reveals a unique
(DFG-out) conformation and enables inhibitor design. J. Biol. Chem. 284,
13193–13201.
Hanks, S.K., and Hunter, T. (1995). Protein kinases 6. The eukaryotic protein
kinase superfamily: kinase (catalytic) domain structure and classification.
FASEB J. 9, 576–596.
Hodous, B.L., Geuns-Meyer, S.D., Hughes, P.E., Albrecht, B.K., Bellon, S.,
Bready, J., Caenepeel, S., Cee, V.J., Chaffee, S.C., Coxon, A., et al. (2007).
Evolution of a highly selective and potent 2-(pyridin-2-yl)-1,3,5-triazine Tie-2
kinase inhibitor. J. Med. Chem. 50, 611–626.
SUPPLEMENTAL DATA
Supplemental Data include Supplemental Experimental Procedures, three
figures, and one table and can be found with this article online at doi:10.1016/
j.chembiol.2010.01.008.
Huang, W.S., Zhu, X., Wang, Y., Azam, M., Wen, D., Sundaramoorthi, R.,
Thomas, R.M., Liu, S., Banda, G., Lentini, S.P., et al. (2009). 9-(Arenethenyl)-
purines as dual Src/Abl kinase inhibitors targeting the inactive conformation:
design, synthesis, and biological evaluation. J. Med. Chem. 52, 4743–4756.
ACKNOWLEDGMENTS
Huse, M., and Kuriyan, J. (2002). The conformational plasticity of protein
kinases. Cell 109, 275–282.
We thank J. Kuriyan (University of California, Berkeley) and M. Seeliger (State
University of New York, Stony Brook) for providing expression plasmids for
ABL (cat), SRC (cat), HCK (cat), HCK (3D), LCK (3D), SRC (3D), and ABL
(3D). We thank S. Knapp (Structural Genomics Consortium) for providing
expression plasmids for CLK1, STK16, EPHA3, PAK4, PAK5, MAP3K5, and
LOK. This work was supported by the National Institute of General Medical
Science (R01GM086858 to D.J.M.).
Iacob, R.E., Pene-Dumitrescu, T., Zhang, J., Gray, N.S., Smithgall, T.E., and
Engen, J.R. (2009). Conformational disturbance in Abl kinase upon mutation
and deregulation. Proc. Natl. Acad. Sci. USA 106, 1386–1391.
Jacobs, M.D., Caron, P.R., and Hare, B.J. (2008). Classifying protein kinase
structures guides use of ligand-selectivity profiles to predict inactive confor-
mations: structure of lck/imatinib complex. Proteins 70, 1451–1460.
Chemistry & Biology 17, 195–206, February 26, 2010 ª2010 Elsevier Ltd All rights reserved 205

Chemistry & Biology
Affinity Reagents for Inactive Protein Kinases
Jeffrey, P.D., Russo, A.A., Polyak, K., Gibbs, E., Hurwitz, J., Massague, J., and
Pavletich, N.P. (1995). Mechanism of CDK activation revealed by the structure
of a cyclinA-CDK2 complex. Nature 376, 313–320.
Schroeder, G.M., An, Y.M., Cai, Z.W., Chen, X.T., Clark, C., Cornelius, L.A.M.,
Dai, J., Gullo-Brown, J., Gupta, A., Henley, B., et al. (2009). Discovery of
N-(4-(2-Amino-3-chloropyridin-4-yloxy)-3-fluorophenyl)-4-ethoxy-1-(4-fluoro-
phenyl)-2-oxo-1,2-dihydropyridine-3-carboxamide (BMS-777607), a selective
and orally efficacious inhibitor of the Met kinase superfamily. J. Med. Chem.
52, 1251–1254.
Kannan, N., Taylor, S.S., Zhai, Y., Venter, J.C., and Manning, G. (2007). Struc-
tural and functional diversity of the microbial kinome. PLoS Biol. 5, e17.
Karaman, M.W., Herrgard, S., Treiber, D.K., Gallant, P., Atteridge, C.E., Camp-
bell, B.T., Chan, K.W., Ciceri, P., Davis, M.I., Edeen, P.T., et al. (2008). A quan-
titative analysis of kinase inhibitor selectivity. Nat. Biotechnol. 26, 127–132.
Seeliger, M.A., Nagar, B., Frank, F., Cao, X., Henderson, M.N., and Kuriyan, J.
(2007). c-Src binds to the cancer drug imatinib with an inactive Abl/c-Kit
conformation and
299–311.
a distributed thermodynamic penalty. Structure 15,
Kornev, A.P., Haste, N.M., Taylor, S.S., and Eyck, L.F. (2006). Surface compar-
ison of active and inactive protein kinases identifies a conserved activation
mechanism. Proc. Natl. Acad. Sci. USA 103, 17783–17788.
Seeliger, M.A., Ranjitkar, P., Kasap, C., Shan, Y., Shaw, D.E., Shah, N.P., Kur-
iyan, J., and Maly, D.J. (2009). Equally potent inhibition of c-Src and Abl by
compounds that recognize inactive kinase conformations. Cancer Res. 69,
2384–2392.
Liu, Y., and Gray, N.S. (2006). Rational design of inhibitors that bind to inactive
kinase conformations. Nat. Chem. Biol. 2, 358–364.
Manning, G., Whyte, D.B., Martinez, R., Hunter, T., and Sudarsanam, S. (2002).
The protein kinase complement of the human genome. Science 298, 1912–
1934.
Simard, J.R., Getlik, M., Gru¨ tter, C., Pawar, V., Wulfert, S., Rabiller, M., and
Rauh, D. (2009a). Development of a fluorescent-tagged kinase assay system
for the detection and characterization of allosteric kinase inhibitors. J. Am.
Chem. Soc. 131, 13286–13296.
Mol, C.D., Dougan, D.R., Schneider, T.R., Skene, R.J., Kraus, M.L., Scheibe,
D.N., Snell, G.P., Zou, H., Sang, B.C., and Wilson, K.P. (2004). Structural basis
for the autoinhibition and STI-571 inhibition of c-Kit tyrosine kinase. J. Biol.
Chem. 279, 31655–31663.
Simard, J.R., Kluter, S., Gru¨ tter, C., Getlik, M., Rabiller, M., Rode, H.B., and
Rauh, D. (2009b). A new screening assay for allosteric inhibitors of cSrc.
Nat. Chem. Biol. 5, 394–396.
Nagar, B., Bornmann, W.G., Pellicena, P., Schindler, T., Veach, D.R., Miller,
W.T., Clarkson, B., and Kuriyan, J. (2002). Crystal structures of the kinase
domain of c-Abl in complex with the small molecule inhibitors PD173955
and imatinib (STI-571). Cancer Res. 62, 4236–4243.
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., et al. (2005).
Prevention of MKK6-dependent activation by binding to p38alpha MAP
kinase. Biochemistry 44, 16475–16490.
O’Hare, T., Shakespeare, W.C., Zhu, X., Eide, C.A., Rivera, V.M., Wang, F.,
Adrian, L.T., Zhou, T., Huang, W.S., Xu, Q., et al. (2009). AP24534, a pan-
BCR-ABL inhibitor for chronic myeloid leukemia, potently inhibits the T315I
mutant and overcomes mutation-based resistance. Cancer Cell 16, 401–412.
Taylor, S.S., Bubis, J., Toner-Webb, J., Saraswat, L.D., First, E.A., Buechler,
J.A., Knighton, D.R., and Sowadski, J. (1988). CAMP-dependent protein
kinase: prototype for a family of enzymes. FASEB J. 2, 2677–2685.
Okram, B., Nagle, A., Adrian, F.J., Lee, C., Ren, P., Wang, X., Sim, T., Xie, Y.,
Wang, X., Xia, G., et al. (2006). A general strategy for creating ‘‘inactive-confor-
mation’’ abl inhibitors. Chem. Biol. 13, 779–786.
Wan, P.T., Garnett, M.J., Roe, S.M., Lee, S., Niculescu-Duvaz, D., Good, V.M.,
Jones, C.M., Marshall, C.J., Springer, C.J., Barford, D., and Marais, R. (2004).
Mechanism of activation of the RAF-ERK signaling pathway by oncogenic
mutations of B-RAF. Cell 116, 855–867.
Pargellis, C., Tong, L., Churchill, L., Cirillo, P.F., Gilmore, T., Graham, A.G.,
Grob, P.M., Hickey, E.R., Moss, N., Pav, S., et al. (2002). Inhibition of p38
MAP kinase by utilizing a novel allosteric binding site. Nat. Struct. Biol. 9,
268–272.
Weisberg, E., Manley, P.W., Breitenstein, W., Bruggen, J., Cowan-Jacob,
S.W., Ray, A., Huntly, B., Fabbro, D., Fendrich, G., Hall-Meyers, E., et al.
(2005). Characterization of AMN107, a selective inhibitor of native and mutant
Bcr-Abl. Cancer Cell 7, 129–141.
Pellicena, P., and Kuriyan, J. (2006). Protein-protein interactions in the allo-
steric regulation of protein kinases. Curr. Opin. Struct. Biol. 16, 702–709.
Perera, B.G., and Maly, D.J. (2008). Design, synthesis and characterization of
Wissing, J., Jansch, L., Nimtz, M., Dieterich, G., Hornberger, R., Keri, G., Weh-
land, J., and Daub, H. (2007). Proteomics analysis of protein kinases by target
class-selective prefractionation and tandem mass spectrometry. Mol. Cell.
Proteomics 6, 537–547.
‘‘clickable’’ 4-anilinoquinazoline kinase inhibitors. Mol. Biosyst. 4, 542–550.
Quintas-Cardama, A., Kantarjian, H., and Cortes, J. (2007). Flying under
the radar: the new wave of BCR-ABL inhibitors. Nat. Rev. Drug Discov. 6,
834–848.
Schindler, T., Bornmann, W., Pellicena, P., Miller, W.T., Clarkson, B., and Kur-
iyan, J. (2000). Structural mechanism for STI-571 inhibition of abelson tyrosine
kinase. Science 289, 1938–1942.
Wong, L., Jennings, P.A., and Adams, J.A. (2004). Communication pathways
between the nucleotide pocket and distal regulatory sites in protein kinases.
Acc. Chem. Res. 37, 304–311.
206 Chemistry & Biology 17, 195–206, February 26, 2010 ª2010 Elsevier Ltd All rights reserved