The design, synthesis, and evaluation of 8 hybrid DFG-out allosteric kinase inhibitors: A structural analysis of the binding interactions of Gleevec , Nexavar, and BIRB-796

Article abstract of DOI:10.1016/j.bmc.2010.05.063

The majority of kinase inhibitors developed to date are competitive inhibitors that target the ATP binding site; however, recent crystal structures of Gleevec (imatinib mesylate, STI571, PDB: 1IEP), Nexavar (Sorafenib tosylate, BAY 43-9006, PDB: 1UWJ), and BIRB-796 (PDB: 1KV2) have revealed a secondary binding site adjacent to the ATP binding site known as the DFG-out allosteric binding site. The recent successes of Gleevec and Nexavar for the treatment of chronic myeloid leukemia and renal cell carcinoma has generated great interest in the development of other kinase inhibitors that target this secondary binding site. Here, we present a structural comparison of the important and similar interactions necessary for Gleevec, Nexavar, and BIRB-796 to bind to their respective DFG-out allosteric binding pockets and the selectivity of each with respect to c-Abl, B-Raf, and p38α. A structural analysis of their selectivity profiles has been generated from the synthesis and evaluation of 8 additional DFG-out allosteric inhibitors that were developed directly from fragments of these successful scaffolds.

Full text of DOI:10.1016/j.bmc.2010.05.063

Bioorganic & Medicinal Chemistry 18 (2010) 5738–5748
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
The design, synthesis, and evaluation of 8 hybrid DFG-out allosteric kinase
inhibitors: A structural analysis of the binding interactions of GleevecÒ,
NexavarÒ, and BIRB-796
Justin Dietrich a,d, , Christopher Hulme a,b,c,d, Laurence H. Hurley a,b,c
*
a Department of Pharmacology and Toxicology, College of Pharmacy, The University of Arizona, Tucson, AZ 85719, United States
b Arizona Cancer Center, University of Arizona, 1515 N. Campbell Avenue, Tucson, AZ 85724, United States
c BIO5 Institute, University of Arizona, 1657 E. Helen Street, Tucson, AZ 85721, United States
d Southwest Comprehensive Center for Drug Discovery and Development, College of Pharmacy, The University of Arizona, Tucson, AZ 85719, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
The majority of kinase inhibitors developed to date are competitive inhibitors that target the ATP binding
site; however, recent crystal structures of GleevecÒ (imatinib mesylate, STI571, PDB: 1IEP), NexavarÒ
(Sorafenib tosylate, BAY 43-9006, PDB: 1UWJ), and BIRB-796 (PDB: 1KV2) have revealed a secondary
binding site adjacent to the ATP binding site known as the DFG-out allosteric binding site. The recent suc-
cesses of GleevecÒ and NexavarÒ for the treatment of chronic myeloid leukemia and renal cell carcinoma
has generated great interest in the development of other kinase inhibitors that target this secondary
binding site. Here, we present a structural comparison of the important and similar interactions neces-
sary for GleevecÒ, NexavarÒ, and BIRB-796 to bind to their respective DFG-out allosteric binding pockets
Received 26 March 2010
Revised 20 May 2010
Accepted 23 May 2010
Available online 4 June 2010
Keyword:
Allosteric structure based drug design
kinase type II
and the selectivity of each with respect to c-Abl, B-Raf, and p38a. A structural analysis of their selectivity
profiles has been generated from the synthesis and evaluation of 8 additional DFG-out allosteric inhibi-
tors that were developed directly from fragments of these successful scaffolds.
Published by Elsevier Ltd.
1. Background
the ATP binding pocket (Fig. 1B). Because the phenyl ring of phen-
ylalanine occupies the space usually utilized to bind ATP, the DFG-
GleevecÒ, which was developed as an inhibitor of c-Abl and the
constitutively active fusion protein form Bcr/Abl,1–3 was approved
for the treatment of chronic myeloid leukemia (CML) in 2002.
GleevecÒ was the first compound shown to inhibit a kinase by
binding the DFG-out allosteric site.4 This site is actually just an
inactive conformation of the enzyme characterized by movement
of an AspPheGly (DFG) loop.5 This movement is common to all
out conformation is mutually exclusive with ATP binding and thus
renders the kinase in an inactive state.
DFG-out inhibitors stabilize the DFG-out conformation by plac-
ing a lipophilic group into the voided hydrophobic pocket, result-
ing in an increase in the population of inactive kinase. DFG-out
allosteric inhibitors typically bridge from this site into the hinge
region, the area normally occupied by ATP. Because allosteric site
inhibitors bind an inactive conformation of the enzyme and do
not compete directly with ATP or substrate, they can offer a signif-
icant kinetic advantage over ATP competitive inhibitors.7
DFG-out inhibitors and is depicted from structures of p38
a in
Figure 1. The position of the phenyl ring of phenylalanine deter-
mines whether the kinase is in the active or inactive conformation.
In the active conformation (DFG-in), which the phosphorylated
protein frequents approximately 80% of the time, the phenyl ring
is contained in a hydrophobic pocket6,7 (Fig. 1A). In this conforma-
tion, the triphosphate of ATP binds by utilizing two magnesium
cofactors and aspartate from the DFG loop. In the inactive confor-
mation (DFG-out), however, rearrangement of the DFG loop allows
the phenyl ring to void the hydrophobic pocket as it rotates around
the C–N bond of aspartate, displacing itself >10 Angstroms and into
NexavarÒ (sorafenib, BAY 43-9006) is a potent inhibitor of
B-Raf, which is currently approved for the treatment of renal cell
carcinoma and is still undergoing multiple clinical trials in thyroid
and liver cancers.8–13 B-Raf kinase is mutated in approximately 7%
of all human cancers with higher occurrence in 66% of human mel-
anomas, 45% of sporadic papillary thyroid cancers, 33% of KRAS
mutated pancreatic cancers, and 15% of sporadic colorectal can-
cers.14–19 In a manner similar to GleevecÒ, this agent inhibits by
binding the inactive conformation of B-Raf characterized by a
DFG loop move. Interestingly, NexavarÒ, like GleevecÒ, is not a spe-
cific inhibitor of its primary target enzyme, but also displays good
inhibitory activity for the kinase domain of the vascular endothe-
* Corresponding author. Tel.: +1 520 626 5622; fax: +1 520 626 4824.
E-mail addresses: dietrich@pharmacy.arizona.edu (J. Dietrich), hurley@pharmacy.
arizona.edu (L.H. Hurley).
0968-0896/$ - see front matter Published by Elsevier Ltd.
doi:10.1016/j.bmc.2010.05.063

J. Dietrich et al. / Bioorg. Med. Chem. 18 (2010) 5738–5748
5739
lial growth factor receptor (VEGFR) and platelet derived growth
factor receptor beta (PDGFRB), which allows this inhibitor to also
promote anti-angiogenic effects much like the effects seen in the
antagonism of the VEGF receptor by AvastinÒ.20 The promising
clinical efficacy of NexavarÒ is attributed to the combination of
inhibiting VEGFR1-2, PDGFR-b, and B-raf.21,22
tion, the lower selectivity site is created when the phenyl alanine
of the DFG loop flips out of its lipophilic pocket during the DFG
loop move. The lower selectivity site is highly conserved among ki-
nases that are approachable through allosteric inhibition. The
upper selectivity site is less conserved and offers a unique position
to build in selectivity. The core scaffold usually binds between the
two selectivity sites and projects lipophilic moieties into the large
hydrophobic pockets. When GleevecÒ binds to Abl, a phenyl ring
occupies space in the lower selectivity pocket and projects a
N-methyl piperazine into the outer rim of the allosteric binding
region (Fig. 2a). NexavarÒ has a phenyl ring which is substituted
with a trifluoromethyl group at the meta position that projects into
the lower selectivity site (Fig. 2b). A para-chloro substituent also
adds hydrophobicity to the phenyl ring and positively interacts
with lower hydrophobic selectivity site when NexavarÒ binds.
BIRB-796 has a pyrazole scaffold that places a lipophilic t-butyl
group into the lower selectivity site and a tolyl ring into the upper
selectivity site (Fig. 2c). Of the three molecules, BIRB-796 is the
only one to utilize the upper selectivity site.
BIRB-796 is a highly potent inhibitor of p38a 23
a serine/threo-
,
nine mitogen activated protein kinase (MAPK) that is usually asso-
ciated with inflammation because of its role in T-cell proliferation
and cytokine production.24 The p38 pathway, however, factors into
many other cellular events including proliferation, apoptosis, and
cell differentiation.21,22,25–28 BIRB-796, like GleevecÒ and NexavarÒ,
binds a highly lipophilic selectivity site created by movement of
the DFG loop. This inhibitor, developed as a treatment for rheuma-
toid arthritis and Crohn’s disease, was withdrawn from phase III
clinical trials.29 Efforts in the pursuit of this inhibitor, however,
yielded some very pertinent structural data that parallels other
type 2 inhibitors. Kinetic data generated during the discovery of
BIRB-796 established the slow binding kinetics of DFG-out inhibi-
tors7 and also highlighted that some allosteric inhibitors can dis-
play the ability to induce a conformation in which both the
activated enzyme can be potently inhibited and the unphosphory-
lated enzyme can not be activated by phosphorylation on the acti-
vation loop.30
2.3. Hydrophobic interactions in the gatekeeper region
The third interaction site is called the ‘gatekeeper region’ which
is close to a conserved lysine side chain that is normally involved in
triphosphate binding with ATP. As a result of the DFG loop move-
ment, the phenylalanine side chain of this loop typically closes off
the gatekeeper region forming a distinct hydrophobic binding site.
An aromatic ring spans this region in all three inhibitors. GleevecÒ
(Fig. 2a) and NexavarÒ (Fig. 2b) utilize a phenyl ring and BIRB-796
utilizes a larger naphthyl ring to bind the gatekeeper region
(Fig. 2c).
2. Structural requirements for allosteric inhibitors
The three DFG-out allosteric kinase inhibitors described share a
basic architecture for construction that can be defined by four key
interactions (Fig. 2).
2.1. Hydrogen-bonding interactions with the core scaffold
2.4. Hydrogen-bonding interactions with the hinge region
A critical inhibitor interacting group is the core or scaffold,
which establishes a bridging hydrogen-bond network between a
conserved glutamate side chain and the amide N–H from the
aspartate involved in the DFG loop move (Fig. 2). In GleevecÒ, this
core scaffold is a benzamide (Fig. 2a), in NexavarÒ a phenyl urea
(Fig. 2b), and in BIRB 796 it is a pyrazole urea (Fig. 2c). The carbonyl
from the urea or amide accepts a hydrogen-bond from the back-
bone NH of aspartate and the NH from the urea or amide forms a
hydrogen-bond to the side chain of the conserved glutamate.
All three inhibitors bind to the hinge region where the adenine
ring of ATP normally binds and accepts a hydrogen-bond from the
hinge region. In GleevecÒ a 3-pyridyl ring is used (Fig. 2a), in Nexa-
varÒ a N-methyl-4-picolinamide (Fig. 2b), and in BIRB-796 a mor-
pholino group (Fig. 2c).
When evaluating the crystal structure of these three inhibitors
bound to their target enzyme, it is not obvious where selectivity
can be achieved because there are many similar binding interac-
tions. This point is exasperated in a recent report by Namboodiri
et al. in which the authors report crystal structures in which Glee-
2.2. Hydrophobic interactions in the allosteric binding region
vecÒ and NexavarÒ are bound to p38a 31
.
In this study, we report
the enzymatic selectivity of GleevecÒ, NexavarÒ, and BIRB-796Ò
with respect to c-Abl, B-Raf, and p38 . Although the selectivity
Hydrophobic interactions are made in the allosteric binding
region within two selectivity sites. In the depicted binding orienta-
a
Figure 1. Conformation DFG-in and DFG-out. (A) The DFG-in active state (magenta) of p38a with ATP bound in the ATP pocket. The three phosphates and the side chain of
0
aspartate complex to the Mg2+ cofactor. (B) The phenyl ring of Phe169 in the active state (magenta) rotates around the CN bond of Asp168 and positions the phenyl ring 10 ÅA
away and into the ATP pocket. This movement results in the DFG-out inactive state shown in green.

5740
J. Dietrich et al. / Bioorg. Med. Chem. 18 (2010) 5738–5748
Figure 2. Crystal structures of known allosteric kinase inhibitors. (a) Crystal structure of GleevecÒ (imatinib, STI571) bound Abl kinase (PDB: 1IEP). (b) Crystal structure of
Sorafenib (NexavarÒ, BAY 43-9006) bound to B-Raf kinase (PDB: 1UWJ). (c) BIRB-796 bound to p38 kinase (PDB: 1KV2. (d) Overlap of GleevecÒ (green), sorafenib (mauve),
a
and BIRB-796 (teal) in their bound conformations. (1) Hydrogen-bonding interactions with the core scaffold. (2) Hydrophobic interactions in the allosteric binding region. (3)
Hydrophobic interactions in the gatekeeper region. (4) Hydrogen-bonding interactions with the hinge region.
of these inhibitors has been previously published,32–34 To further
our study and gain additional information, we have broken these
three molecules down into three distinct fragments (Fig. 3). One
fragment consists of the left side of the central core which binds
to the gatekeeper and hinge regions. The second fragment is the
central core, and the third fragment is the right side of the central
core that binds to the selectivity sites in the allosteric binding re-
gion. In order to distinguish which fragment of each molecule con-
tributes to selectivity among the three kinases of interest, we
synthesized 8 hybrid molecules (Fig. 4) that combine the different
fragments of each known inhibitor and evaluated their inhibitory
activity with respect to c-Abl, B-Raf, and p38a.
Previously described methods were used to synthesize Nexa-
varÒ,35 BIRB-796,7 and GleevecÒ.36 Major intermediates from the
three previous synthetic routes were utilized in the synthesis of
6 of the 8 hybrid molecules ( Scheme 1). Important gatekeeper
and hinge region binding amine intermediates include compound
12, 13, and 14 (Scheme 1). The right sides of hybrid molecules were
derived from isocyanate 16 from sorafenib, 5-tert-butyl-2-p-tolyl-
2H-pyrazol-3-yl-amine (17) from BIRB-796, and 4-[(4-methylpi-
perazin-1-yl)methyl]benzoyl chloride dihydrochloride (18) from
GleevecÒ. Urea and amide linkages were made using the standard
coupling methods as were previously described in the syntheses of
Nexavar, BIRB-796, and GleevecÒ.
Figure 3. Breakdown of allosteric kinase inhibitors.
The synthesis of 10 (Scheme 2) is described in Scheme 2. The
amination of 1-bromomethyl-4-nitrobenzene (18) by 4-N-methyl-
piperazine was accomplished in the presence of potassium
carbonate in refluxing acetonitrile to yield 1-methyl-4-(4-nitro-
benzyl)piperazine (20). The nitro group of 20 was then cautiously
reduced via hydrogenation under 30 psi hydrogen gas in the pres-
ence of 10% palladium on carbon in ethanol to give the correspond-
ing amine (21). The urea linkage was assembled by the reaction of
21 with the preformed N-imidazole carboxamide 22 in dichloro-
methane at room temperature to yield the target molecule 10.
The synthesis of compound 11 was accomplished through the
route described in Scheme 3. The potassium-4-nitro-phenoxide
(24) was generated from reaction of nitrophenol in ethanol in the
presence of potassium hydroxide. After isolation of 24, the ether
linkage was incorporated in the molecule by reaction with

J. Dietrich et al. / Bioorg. Med. Chem. 18 (2010) 5738–5748
5741
Gleevec®
Nexavar®
BIRB-796
H
N
H
N
N
N
H
N
H
N
H
H
N
1O
N
H
N
N
N
O
O
N
N
Cl
N
O
N
O
O
N
O
F
F
F
2
3
Gleevavar-Amide
Nexavec
BIRBvec
N
N
H
N
H
N
H
N
H
N
N
N
N
N
H
N
H
N
H
N
H
N
O
N
N
O
N
Cl
O
N
N
O
O
F
F
F
4
6
8
Nexa-796-Naph
Gleev-796
H
N
H
N
N
H
N
BIRBavar
O
N
O
N
O
O
Cl
N
O
O
F
F
H
N
H
N
F
5
7
N
N
H
N
O
N
O
O
9
Gleevavar-Urea
Nexa-796-Phe
H
N
H
N
H
H
N
N
N
O
N
H
N
N
O
N
O
O
Cl
O
O
F
F
10
11
F
Figure 4. Parent and hybrid molecules that were synthesized and evaluated. GleevecÒ is shown in red, Nexavar in blue, and BIRB-796 in green. The hybrid molecules are
depicted as a combination of colors to represent which part came from which parent molecule. The naming scheme names the right side of the molecule first and the left side
second, that is, Gleev-796 (5) has the right side of GleevecÒ and the left side of BIRB-796.
4-(2-chloroethyl)morpholine in refluxing toluene to afford 4-(2-(4-
nitrophenoxy)ethyl)morpholine (25) in modest yield.37 The nitro
group was then reduced via hydrogenation with 10% palladium
on carbon to yield amine 26. The urea linkage was then accom-
plished by reaction of 26 with isocyanate 16 in dichloromethane.
IC50 of 76.2 nM, p38
IC50 of 225.9 nM.
a with an IC50 of 84.8 nM, and Abl with an
Although the data in Table 1 clearly shows that these three
inhibitors have very different selectivity profiles, there is not en-
ough data and too many variables to ascertain what moieties in
the molecules are responsible for the selectivity profile of each
inhibitor. The enzymatic inhibition profiles of the hybrid mole-
cules, however, can more clearly decipher where the selectivity
is coming from. It is possible to ascertain which portion of the mol-
ecule is responsible for the selectivity by making modifications to
each region of the molecule and evaluating which modification re-
sults in a loss of selectivity.
3. Results and discussion
All of the molecules that were synthesized for this study were
evaluated in three radiometric kinase assays: one for p38
for B-Raf, and one for Abl kinase. The p38 and B-Raf assays uti-
lized human recombinant GST-tagged enzymes purchased from
invitrogen, 100 M ATP, and full length recombinant inactive sub-
a, one
a
Since GleevecÒ is the most selective of the three compounds,
hybrid molecules that contain a moiety from GleevecÒ were eval-
uated first (Table 1, molecules with red). When GleevecÒ (1) or
hybrid molecules such as compounds 4, 5, and 10 utilized the
1-benzyl-4-methylpiperazine from the right side of GleevecÒ, they
l
strate kinase since these kinases have the ability to form ternary
structures with their substrate kinase. The Abl kinase assays uti-
lized mouse GST-tagged c-Abl, the peptide substrate Abltide, and
100 lM ATP.
The initial enzymatic inhibition results from GleevecÒ, Nexa-
varÒ, and BIRB-796 (Table 1, compounds 1–3) showed that among
the three allosteric kinase inhibitors, GleevecÒ was the most spe-
cific for its target enzyme c-Abl, exhibiting an IC50 of 10.8 nM
could not inhibit the enzymatic activity of p38
100 concentration—even when the 4-(2-morpholinoeth-
oxy)naphthyl group from the potent p38 inhibitor BIRB-796
a or B-Raf at a
lM
a
was incorporated (5) or the N-methyl-phenoxypicolinamide group
from NexavarÒ was used (4).
and showing no inhibitory activity against either p38
at a 100 M concentration. BIRB-796 showed the most potency
for its target enzyme p38 with an IC50 of 4.0 nM, but also inhib-
a or B-Raf
On the other hand, when the left side of GleevecÒ was used, as
seen in compounds 6 and 8, then the molecules were able to inhibit
l
a
p38
fluoromethyl-para-chlorophenyl ring from NexavarÒ and inhibits
50% of the activity of c-Abl at less than 1 nM, p38 at 18.4 nM,
a and B-Raf. Compound 6 has a urea linkage and the meta-tri-
ited B-Raf with an IC50 of 83 nM and Abl with an IC50 of 14.6 lM.
NexavarÒ was the least selective and showed strong inhibition in
all three kinase assays, inhibiting its target enzyme B-Raf with an
a

5742
J. Dietrich et al. / Bioorg. Med. Chem. 18 (2010) 5738–5748
NH
2
O
NH
2
N
N
O
NH
2
O
NH
2HCl
N
12
13
NH
N
14
O
N
O
NH
O
N
O
N
NH
4
N
Cl
a
O
HCl
13 or 14
N
15
N HCl
O
NH
N
O
N
5
O
N
NH
O
NH
N
N
NH
O
C
6
N
F C
3
Cl
b
N
Cl
12 14
or
16
F
F
F
NH
O
O
NH
N
7
8
O
F C
Cl
3
NH
O
N
NH
NH
N
N
N
H N
2
N
N
c
17
N
12 or 13
NH
O
O
NH
N
N
N
NH
9
O
Scheme 1. Synthesis of hybrid molecules. Reagents and conditions: (a) 13 or 14, pyridine, 50 °C, 12 h; (b) 12 or 14, DCM, TEA, 12 h; (c) phosgene, 4 h, then 12 or 13, 12 h.
and B-Raf at 180 nM. Compound 8 has a urea linkage and the
5-tert-butyl-2-p-tolyl-2H-pyrazol-3-yl ring from BIRB-796 and
and not B-Raf or p38a comes from the 1-benzyl-4-methylpipera-
zine moiety that binds in the allosteric binding region of c-Abl. This
observation has previously been reported by Baldwin et al. in
which compounds containing a N-methyl-piperazine did not show
inhibits c-Abl with an IC50 of 8.6 nM, p38
a with an IC50 of
183.9 nM, and B-Raf with an IC50 of 413.9 nM. From this data, it
is apparent that the selectivity of GleevecÒ to only inhibit c-Abl
any inhibition of p38a
activity.38

J. Dietrich et al. / Bioorg. Med. Chem. 18 (2010) 5738–5748
5743
NO2
NH2
NO2
H
N
b
a
N
N
N
Br
N
N
21
20
19
18
NH
H
N
H
N
c
O
N
N
O
O
O
O
N
N
O
N
H
N
N
10
22
NH
Scheme 2. Synthesis of compound 10. Reagents and conditions: (a) K2CO3, MeCN, reflux, 18 h, 45%; (b) 30 psi H2 gas, 10% Pd/c, EtOH, 2 h, 99%; (c) 21, DCM, rt, 24 h, 78%.
O
NO2
NH2
N
NO2
NO2
Cl
c
a
b
O
O
N
OH
23
OK
24
N
25
26
O
N
O
O
C
H
H
N
N
N
O
d
O
Cl
O
Cl
F
F
F
F
F
F
11
16
Scheme 3. Synthesis of compound 11. Reagents and conditions: (a) KOH, EtOH, rt, 2 h, 90%; (b) 4-(2-chloroethyl)-morpholine, toluene, reflux, 24 h, 54%; (c) 50 psi H2, 10% Pd/c,
EtOH, 1 h, 99%; (d) DCM, rt, 18 h, 65%.
If this hypothesis is true, it is expected to find a structural differ-
ence in the crystal structures of c-Abl, p38, and B-Raf that can
explain the molecular reason for this selectivity. In GleevecÒ, the
methyl piperazine sits deeply into the selectivity site and in its bio-
logically prevalent protonated form makes two hydrogen-bonds
with the backbone carbonyls of isoleucine 360 and histidine 361
(Fig. 5). Two methylene carbons in the piperazine ring make van
der Waals interactions with the side chain benzyl methylene of
phenylalanine 359. As depicted by the overlay of B-Raf with
c-Abl (Fig. 5), the selectivity site residues are highly homologous.
8.6 nM and 62.6 nM. The inhibitory activity of these molecules is
much lower for B-Raf. Compound 8 inhibits B-Raf with an IC50 of
413.9 nM and compound
9 inhibits B-Raf with an IC50 of
236.7 nM. These results suggest that the large 5-tert-butyl-2-p-to-
lyl-2H-pyrazol-3-yl moiety from BIRB-796 binds unfavorably to
the allosteric binding region of B-Raf and a large hydrophobic moi-
ety engineered into molecules to bind this region can be used to
make a p38a inhibitor selective over B-Raf because the overall size
of the hydrophobic selectivity sites in B-Raf is smaller.
When the left side of molecules utilized the 4-(2-morpholino-
ethoxy)naphthyl group from BIRB-796 (Table 1, compounds 3, 5,
and 7) and the right sides of the molecules were changed, it was
apparent that the 4-(2-morpholinoethoxy)naphthyl group does
not bind favorably with c-Abl. In compounds 3, 5, and 7 (Table 1),
the IC50’s for c-Abl are all in the micromolar range. When comparing
the p38 and B-Raf enzymatic inhibition data, the 4-(2-morpholino-
ethoxy)naphthyl group can make favorable interactions in both p38
and B-Raf. Compound 7 is a very potent inhibitor of both p38 and B-
Raf with IC50’s for each enzyme below 1 nM.
To structurally evaluate the differences in the three kinases at
the gatekeeper region, we superimposed the three crystal struc-
tures on top of each other and evaluated this portion of the mole-
cule and found that the naphthyl ring from BIRB-796 pushes very
deep into the gate keeper region. At the deepest part of this region
In B-Raf and p38
a, the analogous position to phenylalanine 359
in c-Abl is instead an isoleucine (p38
a
ILE146, B-Raf ILE571), which
directs the isolated methyl group into the analogous space pipera-
zine occupies in c-Abl and blocks the binding pocket. This finding is
further supported by the resistance profiles of patients that are
prescribed GleevecÒ. A rare but very serious single nucleotide sub-
stitution at codon 359 that replaces phenylalanine with an isoleu-
cine can occur in patients that receive GleevecÒ therapy and this
single mutation results in complete resistance to GleevecÒ chemo-
therapy.39 The same residue that is responsible for the selectivity
of GleevecÒ over p38 and B-Raf is an Achilles’ heel, that if mutated,
will also knock out inhibition of c-Abl.
The same type of analysis can be performed with BIRB-796 if
the right side of BIRB-796 is held constant with the 5-tert-butyl-
2-p-tolyl-2H-pyrazol-3-yl group from BIRB-796 (Table 1, com-
pounds 3, 8 and 9) and the left side is modified. Compounds 8
and 9 display potent inhibitory activity for c-Abl with IC50’s of
in p38a, leucine 104 makes hydrophobic interactions with the
naphthyl ring (Fig. 6). In c-Abl, however, this position is instead iso-
leucine 313 which decreases the size of the hydrophobic pocket by

5744
J. Dietrich et al. / Bioorg. Med. Chem. 18 (2010) 5738–5748
Table 1
Complete enzymatic results (results were calculated from the average of n P 2 independent experiments performed in duplicate with R2 >0.93)
IC50 Abl
nM
IC50 B-Raf
nM
IC50 p38a
nM
M.W.
493
cLog P
N
N
H
N
H
N
N
GleevecÒ
NexavarÒ
1
3.83
10.8
>100,000
76.2
>100,000
84.8
N
N
O
O
H
N
H
N
N
H
N
2
464
527
3.76
5.79
225.9
Cl
O
O
F
F
F
H
H
BIRB-796 3
14600.0
83.4
4.0
N
N
N
O
N
N
O
O
N
H
N
N
N
H
N
GleevAvar 4
Gleev-796 5
459
488
2.30
3.01
>30,000
>30,000
>100,000
>100,000
>100,000
>100,000
O
O
O
N
H
N
N
O
N
O
O
N
H
N
H
H
N
N
N
NexaVec 6
498
493
5.28
4.46
<1
180.1
18.4
<1
N
O
Cl
F
F
F
F
H
H
N
N
O
N
O
Nexa-796 7
12930.0
<1
O
Cl
F
F
N
H
N
H
H
N
BIRBvec 8
BIRBavar 9
532
498
6.61
5.08
8.6
413.9
236.7
189.3
N
N
N
N
N
O
O
H
N
H
N
62.6
<1
N
N
N
H
N
O
O
O
H
N
H
N
N
N
H
N
N
Gleevavar-Urea 10
Nexa-796-Phe 11
474
443
3.61
5.10
572.4
244.0
>100,000
92.3
>100,000
105.5
O
O
H
H
N
N
O
N
O
O
Cl
F
F
F
projecting the isolated methyl group into the hydrophobic pocket.
Consequentially, the gatekeeper region of c-Abl is responsible for
the unfavorable binding of BIRB-796 to c-Abl.
evidence that the naphthyl ring was too large to fit into the
gatekeeper region of c-Abl. Compound 7 (Table 1) displayed
subnanomolar potency for both p38
a and B-Raf and has a naph-
To further investigate this hypothesis, compound 11 (Table 1)
was synthesized, which contained a morpholinoethoxy phenyl
group instead of the morpholinoethoxy naphthyl group. Com-
pound 11 is identical to compound 7 except for the naphthyl ring
in compound 7 is changed to a smaller phenyl ring. Compound 11
was able to rescue c-Abl activity (IC50 = 244.0 nM) and provided
thyl ring in the gatekeeper region, however, when the smaller
phenyl ring was used in compound 11, the potency decreased
(IC50BRaf = 92.3 nM, IC50p38 = 105.5 nM) indicating that there are
favorable hydrophobic binding interactions when a naphthyl
ring is utilized to bind to the gatekeeper region of B-Raf and
p38a.

J. Dietrich et al. / Bioorg. Med. Chem. 18 (2010) 5738–5748
5745
It was surprising that compound 4 did not show any inhibition
of c-Abl at 30 M. We hypothesized that this was due to having an
l
amide in the core instead of a urea. Compound 10, which is iden-
tical to compound 4 except for that it has a urea in the core instead
of an amide, was synthesized to explore this hypothesis. When this
compound was tested, it expectedly showed no inhibition of p38
a
and B-Raf at 100 M, but modestly inhibited c-Abl with an IC50 of
l
572.4 nM. When the urea scaffold is used with the N-methyl-
4-phenoxypicolinamide moiety on the left, it can make favorable
binding interactions in the hinge region, but as a consequence of
the longer linkage, the interactions in the allosteric binding region
by the 1-benzyl-4-methylpiperazine from GleevecÒ are then aber-
rantly modified. As discussed earlier and depicted in Figure 5, the
N-methylpiperazine ring sits in a very tight pocket and any modi-
fication at this region can result in a steric interaction that is not
favorable for binding.
Figure 5. Overlap of c-Abl (1IEP) with B-Raf kinase (PDB: 1UWJ). c-Abl and Gleevec
are shown in green. B-Raf kinase is shown in teal. Isoleucine 571 is not compatible
with GleevecÒ binding to B-Raf. A single residue difference at this position is
responsible for the ability of GleevecÒ to select c-Abl over p38
a and B-Raf kinase.
4. Summary and conclusions
The research presented here evaluates the structural overlap of
three DFG-out allosteric kinase inhibitors, GleevecÒ, NexavarÒ, and
BIRB-796, bound to their primary target, C-Abl, B-Raf, and p38a.
These known molecules were broken down into three distinct
regions and 8 hybrid molecules were synthesized that were
derived directly from the scaffolds of these successful inhibitors
(Figs. 3 and 4). Enzymatic inhibition results from the known inhib-
itors and the 8 hybrid DFG-out inhibitors gave insight into the
regions of these molecules and kinases that promote selectivity
or broad range inhibition (Fig. 7).
In the selectivity site, which is available upon the DFG rear-
rangement, the 3-trifluoromethyl-4-chlorophenyl ring found in
NexavarÒ can make favorable interactions in all three kinases.
When this moiety is changed to a larger moiety such as the
5-tert-butyl-2-p-tolyl-2H-pyrazole as found in BIRB-796, this tends
to alleviate B-Raf binding and favors binding to p38a
and GleevecÒ.
When GleevecÒ binds to C-Abl, the selectivity site is occupied
by a 1-benzyl-4-methylpiperazine. In this system, the N-methyl-
piperazine ring binds to the outer edge of the selectivity sites
and donates two hydrogen-bonds to carbonyls from the backbone
of the protein. The carbons of the piperazine ring make a positive
Figure 6. Overlap in the gatekeeper region of cAbl and p38a
. GleevecÒ and c-Abl
(PDB: 1IEP) is shown in green and BIRB-796 and p38 (PDB: 1KV2) is shown in teal.
and unique van der Waal’s interaction with Phe369. In p38
a and
NexavarÒ is the least selective inhibitor of the three. When the
meta-trifluoromethyl-para-chlorophenyl ring is utilized to make
inhibitors (compounds 2, 6, 7, 11) and the left side is modified, it
is possible to make subnanomolar inhibitors for all three kinases
(Table 1). Compound 6 is a subnanomolar inhibitor of c-Abl and
B-Raf, this interaction is not possible because the analogous posi-
tion contains an isoleucine that blocks the pocket in which the
N-methylpiperazine ring of GleevecÒ binds c-Abl.
Placing a large naphthyl ring in the gatekeeper region makes
favorable interactions with p38
a smaller hydrophobic pocket. The naphthyl ring in BIRB-796 is
responsible for the selectivity of p38 over c-Abl.
a and B-Raf, however, Abl contains
compound 7 is a subnanomolar inhibitor of both p38
a and B-Raf.
a
Because this moiety can be used to make potent inhibitors for all
three target kinases, it is prudent to conclude that this moiety
binds well to the allosteric binding region of all three kinases. As
discussed previously, the data collected for compounds 3, 8, and
9 indicates that B-Raf has a smaller allosteric binding region than
p38 and c-Abl. As a result of this, a B-Raf inhibitor must have a
smaller moiety in this region and consequentially cannot select
B-Raf kinase over p38 or c-Abl.
When the left side of hybrid molecules utilized the N-methyl-4-
phenoxy-picolinamide moiety from NexavarÒ (compounds 2, 4, 9,
10) and the right side is modified, it is possible to make potent
inhibitors for all three targets as well. NexavarÒ itself is a potent
inhibitor of both B-Raf (IC50 = 76.2 nM) and p38 (IC50 = 84.8 nM)
and a slightly less potent inhibitor of c-Abl (IC50 = 225.9). Com-
pound 9, which also utilizes the N-methyl-4-phenoxy-picolina-
mide is also a potent inhibitor of all three kinases. The broad
spectrum inhibition of NexavarÒ can be attributed to the ability
of both the right side and left side of NexavarÒ to bind with high
affinity to all three kinases.
Aside from the structural data that was gathered from synthe-
sizing and evaluating these molecules, we have also created 8
new chemical entities that have varying attributes such as novel
selectivity profiles, increased potency, and/or favorable drug-like
properties. Compounds 6, 8, and 9 are all potent inhibitors of
c-Abl and structurally represent novel compounds that could pre-
dictably inhibit mutated forms of BCR-Abl that are resistant to
GleevecÒ inhibition. These studies are ongoing and will be reported
separately in due course. Compound 7 (Table 1) is a novel, subn-
anomolar inhibitor of p38 and B-Raf kinase and could be explored
for its utility in targeting B-Raf mutated cancers.
This research highlights that selectivity in kinase inhibitors is a
reachable goal if inhibitors are engineered to exploit minor differ-
ences among kinases. A very simple one residue difference in the
allosteric binding region of c-Abl allows GleevecÒ to be selective
for c-Abl over p38
the gatekeeper region of p38 allows BIRB-796 to be selective for
p38 over c-Abl. Finding unique differences among kinases and
a and B-Raf. Another one residue difference in
a

5746
J. Dietrich et al. / Bioorg. Med. Chem. 18 (2010) 5738–5748
Figure 7. Schematic of specific binding interactions in GleevecÒ, NexavarÒ, and BIRB-796. Red represents atoms from GleevecÒ, blue represents atoms from NexavarÒ, and
green represents atoms from BIRB-796.
kinase inhibitors that render selectivity is only the first battle. As it
becomes more possible to design molecules with a desired kinase
profile, the next stage will have to address which profile is the
most desirable.
cipitate formed that was collected by vacuum filtration and puri-
fied by flash chromatography through a silica gel column on a
Biotage Sp1 system (3–8% MeOH/DCM) to yield 106 mg
4
(0.231 mmol, 64.5% yield). 1H NMR (300 MHz, DMSO-d6): d 10.57
(s, 1H), 8.83 (q, 1H, J = 4.8 Hz), 8.52 (d, 1H, J = 5.5 Hz), 8.05 (d,
2H, J = 8.2 Hz), 7.95 (d, 2H, J = 9.0 Hz), 7.82 (d, 2H, J = 7.1 Hz),
7.41 (d, 1H, J = 2.5 Hz), 7.26 (d, 2H, J = 8.5 Hz), 7.19 (q, 1H,
J = 1.5 Hz), 4.44 (s, 2H), 3.61 (br, 8H), 2.81 (s, 3H), 2.78 (d, 3H,
5. Experimental procedures
5.1. Chemistry
J = 4.6 Hz). 13C NMR (75 MHz, DMSO-d6):
d 166.86, 165.83,
1H NMR and 13C NMR spectra were recorded on Bruker Avance
spectrometer at 300 and 75 MHz, respectively, with TMS as an
164.47, 153.08, 151.15, 149.66, 137.78, 136.33, 132.14, 128.96,
125.26, 123.11 (2c), 122.17 (2C), 115.02, 109.64, 65.87, 58.40
(2C), 51.29 (2C), 48.74, 26.90. HRMS (ESI) (M+1): 460.2343 calcd
for C26H29N5O3 (M+1) = 460.2270.
internal standard. HRMS experiments were performed on
a
Q-TOF-2TM (Micromass). TLC was performed with Merck 60
F254 silica gel plates. Column chromatography was performed
using a Biotage SP1 system with manually loaded 12-, 25-, or
40-g columns. The purity of the target compounds was determined
to be >95% by LC/MS on a Series 1100 LC/MSD (Agilent Technolo-
gies, Palo Alto, CA) equipped with a vacuum de-gasser (G1322A),
a binary pump (1312A), an autosampler (G1313A), a thermostated
column compartment (G1316A), and a mass selective detector
(G1946A) supplied with atmospheric pressure ionization electro-
5.2.2. 4-((4-Methylpiperazin-1-yl)methyl)-N-(4-(2-morpholino-
ethoxy)-naphthalen-1-yl)benzamide (Gleev-796, 5)
141 mg 4-((4-methylpiperazin-1-yl)methyl)benzoyl chloride
dihydrochloride (0.434 mmol, 1.5 equiv) and 100 mg 15 (0.290
mmol, 1 equiv) were reacted in 5 mL THF for 14 h and then poured
into a separatory funnel that contained 50 mL sodium carbonate
and 50 mL EtOAc. The organic layer was removed and the water
layer was then extracted with another 50 mL of EtOAc. The com-
bined organic layer was dried over MgSO4 and concentrated to
near dryness as a white solid. The product was very hygroscopic
and was stored under nitrogen in a desiccator to prevent decompo-
sition. 1H NMR (300 MHz, DMSO-d6): d 10.17 (s, 1H), 8.19 (dd, 1H,
J = 1.9 Hz, J = 7.7 Hz), 7.43–7.63 (m, 3H), 7.36 (d, 2H, J = 8.4 Hz),
7.97 (d, 2H, J = 8.4 Hz), 4.27 (t, 2H, J = 5.7 Hz), 3.41–3.60 (m, 6H),
2.88 (m, 2H), 2.2–2.6 (br d, 12 H), 2.23 (s, 3H), 2.06 (s, 3H). HRMS
(ESI) (M+1): 489.2860 calcd for C29H36N4O3 (M+1) = 489.2787.
spray (API-ES). LC/MS utilized a ZorbaxÒ C18 SB column (3.5
lm,
4.6 ꢀ 150 mm), MeCN/HOH/HAc (pH 4.8) elution buffer, and
15-min runs. Mass spectra and high resolution mass spectra were
obtained using an ESI method by the Mass Spectroscopy Facility in
the Chemistry Department at the University of Arizona. All
solvents and reagents were purchased from Aldrich, TCI America,
Matrix Scientific, and Ryan Scientific and used without further
purification or drying.
5.2. Preparation of hybrid molecules
5.2.3. 1-(4-Chloro-3-(trifluoromethyl)phenyl)-3-(4-methyl-3-(4-
(pyridin-3-yl)pyrimidin-2-ylamino)phenyl)urea (Nexavec, 6)
5.2.1. N-Methyl-4-(4-(4-((4-methylpiperazin-1-yl)methyl)-
benzamido)phenoxy)picolinamide (Gleevavar-Amide, 4)
117 mg 4-((4-methylpiperazin-1-yl)methyl)benzoyl chloride
dihydrochloride (0.357 mmol, 1 equiv) and 100 mg 15 (0.357
mmol, 1 equiv) was reacted in 5 mL pyridine for 14 h and then
poured onto 50 g ice. The mixture was allowed to cool and a pre-
To
a stirred solution of 100 mg (0.514 mmol, 1.25 equiv)
4-chloro-3-trifluoromethylaniline in 5 mL DCM was added 5 mL
NaHCO3 and the mixture was cooled to 0 °C in an ice bath. Stirring
was stopped and 384
lL (0.617 mmol, 1.5 equiv) of a 20% solution
of phosgene in toluene was added directly to the DCM layer via

J. Dietrich et al. / Bioorg. Med. Chem. 18 (2010) 5738–5748
5747
syringe. Stirring was then continued for 4 h. 113 mg amine 12
(0.409 mmol, 1 equiv) was then added and the reaction was al-
lowed to proceed for 14 h. The reaction was then dissolved in
50 mL EtOAc and subsequently washed with a saturated solution
of sodium bicarbonate and then water. The crude material was
dried onto silica, which was then loaded onto a silica gel column.
The column was run on a Biotage Sp1 system (1–6%, MeOH, CDCl3)
to yield 158 mg 6 (0.317 mmol, 77% yield) as a light yellow solid.
1H NMR (300 MHz, DMSO-d6): d 9.29 (d, 1H, 1.6 Hz), 9.10 (s,
1H), 8.90 (s, 1H), 8.77 (s, 1H), 8.69 (dd, 1H, J = 1.5 Hz, J = 4.7 Hz),
8.47–8.52 (m, 2H), 8.11 (d, 1H, J = 2.0 Hz), 7.83 (s, 1H), 7.58–7.65
(m, 2H), 7.51 (dd, 1H, J = 4.8 Hz, J = 7.9 Hz), 7.43 (d, 1H,
J = 5.2 Hz), 7.15 (s, 1H), 2.21 (s, 3H). 13C NMR (75 MHz, DMSO-
d6): d 162.24, 161.86, 160.37, 153.26, 151.75, 148.61, 140.32,
138.78, 137.97, 135.85, 133.25, 132.81, 131.14, 127.51 (d,
J = 30.7 Hz), 126.70, 124.84, 123.81, 123.67 (q, J = 273.3), 122.96,
117.49 (q, J = 5.9 Hz), 115.97, 115.75, 108.49, 18.38. HRMS (ESI)
(M+1): 499.1295 calc. for C24H18ClF3N6O (M+1) = 498.1183.
126.18, 125.30 (2C), 124.70, 115.23, 115.05, 108.48, 79.93, 32.86,
31.06, 21.46, 18.34. HRMS (ESI) (M+1): 533.2772 calcd for
C31H32N8O (M+1) = 533.2699.
5.2.6. 4-(4-(3-(3-tert-Butyl-1-p-tolyl-1H-pyrazol-5-yl)ureido)-
phenoxy)-N-methylpicolinamide (BIRBavar, 9)
To a stirred solution of 100 mg (0.436 mmol, 1.5 equiv) 28 in
5 mL DCM was added 5 mL NaHCO3 and the mixture was cooled
to 0 °C in an ice bath. Stirring was stopped and 277 lL
(0.523 mmol, 1.5 equiv) of a 20% solution of phosgene in toluene
was added directly to the DCM layer via syringe. Stirring was then
continued for 4 h. 98 mg amine 13 (0.349 mmol) was then added
and the reaction was allowed to proceed for 14 h. The reaction
was added to
a 250 mL separatory funnel containing water
(20 mL) and DCM (100 mL). The organic layer was separated, dried
over MgSO4, and then concentrated in vacuo to yield a yellow solid
that was purified by flash column chromatography on a Biotage
Sp1 system (1–5% MeOH/DCM) to yield 120 mg of 8 (0.241 mmol,
69.0% yield) of an off white solid, which was dried in a vacuum
oven at 60 °C for 6 h.
5.2.4. 1-(4-Chloro-3-(trifluoromethyl)phenyl)-3-(4-(2-morpho-
linoethoxy)-naphthalen-1-yl)urea (Nexa-796-Naph. 7)
1H NMR (300 MHz, DMSO-d6): d 9.16 (s, 1H), 8.85 (q, 1H,
J = 4.8 Hz), 8.52 (d, 1H, J = 5.5 Hz), 8.37 (s, 1H), 7.53 (d, 2H,
J = 5.6 Hz), 7.33–7.43 (m, 5H), 7.12–7.16 (m, 3H), 6.37 (s, 1H),
2.79 (d, 3H, J = 4.8 Hz), 2.38 (s, 3H), 1.28 (s, 9H). 13C NMR
(75 MHz, DMSO-d6): d 166.85, 164.63, 161.39, 153.28, 152.47,
151.21, 148.45, 138.17, 137.94, 137.67, 136.93, 130.57 (2C),
125.24 (2C), 122.35, 120.71, 114.85, 109.52, 96.00, 32.88, 31.08
(3C), 26.86, 21.47. HRMS (ESI) (M+1): 499.2452 calcd for
To
a stirred solution of 100 mg (0.514 mmol, 1.25 equiv)
4-chloro-3-trifluoromethylaniline in 5 mL DCM was added 5 mL
NaHCO3 and the mixture was cooled to 0 °C in an ice bath. Stirring
was stopped and 384 lL (0.617 mmol, 1.5 equiv) of a 20% solution
of phosgene in toluene was added directly to the DCM layer via
syringe. Stirring was then continued for 4 h. 141 mg amine 14
(0.409 mmol, 1 equiv) was then added and the reaction was al-
lowed to proceed for 14 h. The product precipitated from the reac-
tion as a pink solid and was collected by suction filtration and
subsequently washed with water and DCM. The light pink material
was then dried in a vacuum oven at 60 °C for 6 h to yield 123 mg of
7 (60.9% yield). 1H NMR (300 MHz, DMSO-d6): d 9.34 (s, 1H), 8.62
(s, 1H), 8.20 (dd, 1H, J = 1.3 Hz, J = 8.3 Hz), 8.14 (d, 1H, J = 2.5 Hz),
7.99 (d, 1H, J = 7.9 Hz), 4.28 (t, 2H, J = 5.7 Hz), 3.60 (m, 4H), 2.86
(t, 2H, J = 5.7 Hz), 2.56 (m, 4H). 13C NMR (75 MHz, DMSO-d6): 13C
C28H30N6O3 (M+1) = 499.2379.
5.2.7. N-Methyl-4-(4-(3-(4-((4-methylpiperazin-1-yl)methyl)-
phenyl)-ureido)phenoxy)picolinamide (Gleevavar-Urea, 10)
A solution of 300 mg (1.072 mmol, 1 equiv) 12 in DCM was stir-
red under argon at room temperature followed by the addition of
174 mg (1.072 mmol, 1 equiv) CDI. The reaction was stirred for
16 h and then a solution of 176 mg of 21 (0.858 mmol, 0.8 equiv)
in 5 mL DCM was added dropwise. The reaction was stirred for
an additional 24 h upon which time it was poured into 100 mL
DCM and then washed with water and brine. The crude material
was purified on an Biotage Sp1 system with a silica gel column that
was neutralized with a 1% solution of triethylamine in CHCl3. The
column was run with a 1–6% MeOH/CHCl3 gradient that also con-
tained 1% triethylamine and provided 176 mg of 10 (0.371 mmol,
34.6% yield) which was dried in a vacuum oven at 60 °C.
NMR (75 MHz, DMSO-d6):
d 154.40, 152.11, 140.59, 132.83,
129.76, 127.35, 127.32 (q, J = 30.7 Hz), 127.20, 126.26, 126.09,
123.71 (q, J = 272.8), 123.67, 122.91, 122.86, 122.79, 122.25,
117.41 (q, J = 5.5 Hz), 67.10, 67.05, 57.87, 54.52. HRMS (ESI)
(M+1): 494.1453 calcd for C24H23ClF3N3O3 (M+1) = 494.1380.
5.2.5. 1-(3-tert-Butyl-1-p-tolyl-1H-pyrazol-5-yl)-3-(4-methyl-3-
(4-(pyridin-3-yl)pyrimidin-2-ylamino)phenyl)urea (BIRBvec, 8)
To a stirred solution of 100 mg (0.436 mmol, 1.5 equiv) 28 in
5 mL DCM was added 5 mL NaHCO3 and the mixture was cooled
1H NMR (300 MHz, CDCl3): d 8.39 (d, 1H, J = 5.6 Hz), 8.25 (q, 1H,
J = 2.2 Hz), 8.17 (s, 1H), 8.07 (s, 1H), 7.60 (d, 1H, J = 2.5 Hz), 7.19–
7.37 (m, 7H), 7.03 (dd, 1H, J = 2.5 Hz, J = 5.6 Hz), 6.94 (d, 2H,
J = 7.9 Hz), 3.43 (s, 2H), 3.03 (d, 3H, J = 5.1 Hz), 2.30–2.70 (br m,
8H), 2.29 (s, 3H). 13C NMR (75 MHz, CDCl3-d6): d 167.21, 165.67,
153.87, 151.90, 150.34, 148.54, 138.16, 137.50, 132.96 (2C),
130.30, 121.73 (2C), 121.36 (2C), 120.04 (2C), 115.21, 109.35,
62.87, 55.40 (2C), 53.29 (2C), 46.34, 26.85. HRMS (ESI) (M+1):
475.2452 calcd for C26H30N6O3 (M+1) = 475.2379.
to 0 °C in an ice bath. Stirring was stopped and 277 lL
(0.523 mmol, 1.5 equiv) of a 20% solution of phosgene in toluene
was added directly to the DCM layer via syringe. Stirring was then
continued for 4 h. 97 mg amine 12 (0.349 mmol) was then added
and the reaction was allowed to proceed for 14 h. The reaction
was added to
a 250 mL separatory funnel containing water
(20 mL) and DCM (100 mL). The organic layer was separated, dried
over MgSO4, and then concentrated in vacuo to yield a brown res-
idue that was purified by flash column chromatography on a Bio-
5.2.8. 1-(4-Chloro-3-(trifluoromethyl)phenyl)-3-(4-(2-morpho-
linoethoxy)-phenyl)urea (Nexa-796-Phe, 11)
tage Sp1 system (3–8% MeOH/DCM) to yield 143 mg of
8
(0.231 mmol, 66.2% yield) as a white solid, which was dried in a
vacuum oven at 60 °C for 6 h.
150 mg isocyanate 16 (0.677 mmol, 1.0 equiv) was dissolved in
5 mL DCM. A solution of 149 mg amine 26 in DCM was added drop-
wise to the stirred reaction and was allowed to react for 12 h. The
reaction was dissolved in 100 mL DCM and washed with water,
washed with brine, and then was dried over MgSO4 and concen-
trated to yield a crude yellow oil. The yellow oil was purified
through flash chromatography on a Biotage Sp1 (2–8% MeOH/
DCM) to yield 165 mg of 11 (0.373 mmol, 55.2% yield)
1H NMR (300 MHz, DMSO-d6): d 9.29 (d, 1H, J = 1.6 Hz), 8.95 (s,
1H), 8.90 (s, 1H), 8.85 (s, 1H), 8.69 (dd, 1H, J = 1.6 Hz, J = 4.8 Hz),
8.47–8.52 (m, 2H), 8.31 (s, 1H), 8.29 (s, 1H), 7.82 (d, 1H,
J = 1.9 Hz), 7,52 (m, 1H), 7.43 (d, 1H, J = 5.2 Hz), 7.39 (dd, 2H,
J = 2.0 Hz, J = 6.5 Hz), 7.35 (d, 2H, J = 6.5 Hz), , 6.38 (s, 1H), 2.37 (s,
3H), 2.19 (s, 3H), 1.26 (s, 9H). 13C NMR (75 MHz, DMSO-d6): d
162.39, 161.83, 161.41, 160.33, 152.30, 152.16, 149.02, 138.80,
138.15, 137.70, 136.83, 135.41, 133.05, 131.14, 130.59 (2C),
1H NMR (300 MHz, DMSO-d6): d 9.01 (s, 1H), 8.65 (s, 1H), 8.10
(d, 1H, J = 2.0 Hz), 7.57–7.61 (m, 2H), 7.35 (d, 2H, J = 9.0 Hz), 6.88

5748
J. Dietrich et al. / Bioorg. Med. Chem. 18 (2010) 5738–5748
(d, 2H, J = 9.0 Hz), 4.03 (t, 2H, J = 5.7 Hz), 3.57 (m, 4H), 2.66 (t, 2H,
J = 5.7 Hz), 2.47 (m, 4H): d 154.84, 153.41, 140.44, 133.04, 132.79,
127.52 (q, J = 30.5 Hz), 123.73, 123.69 (q, J = 273.0),122.84, 121.36
(2C), 117.44 (q, J = 6.0 Hz), 115.50 (2C), 67.032 (2C), 66.33, 57.94,
54.51 (2C). HRMS (ESI) (M+1): 444.1296 calcd for C20H21ClF3N3O3
(M+1) = 444.1224.
6.4. Abl enzymatic assay
The Abl kinase assay utilized 1 ng/rxn active recombinant
mouse Abl, residues 27-1123, containing an N-terminal His6 tag
(Millipore, 14-459) and the 100 lM synthetic peptide substrate
Abltide (EAIYAAPFAKKK, Millipore, 12-493). Kinase reactions were
set up as previously described in the p38 enzymatic assay, how-
5.3. Molecular overlap
ever, reactions were quenched with 5
10
lL 3% phosphoric acid.
L of each reaction was then applied to 2 ꢀ 2 cm p81 phospho-
l
Three dimensional crystal structures of GleevecÒ bound to
C-Abl (2HYY), NexavarÒ bound to B-Raf (1UWH), and BIRB-796
bound to p38a (1KV2) were downloaded from Protein Data Bank.
Water molecules were removed from each crystal structure and
monomers of 1KV2 and 1UWH were aligned with 2HYY using Py-
mol v1.1 from Delano Scientific.
cellulose squares (Whatmann) and washed a minimum of five
times for 5 min with 75 mM phosphoric acid. Squares were then
washed with methanol for 2 min and placed in a 24 well Perkin–El-
mer plate with 1 mL of scintillation fluid and counted on a Perkin–
Elmer Top Count scintillation counter.
Acknowledgment
6. Biological evaluation
6.1. General
This research was supported by a grant from the National Foun-
dation for Cancer Research.
References and notes
All assays utilized assay dilution buffer and Mg/ATP cocktail
purchased from Millipore. The assay dilution buffer contained
20 mM MOPS, pH 7.2, 25 mM b-glycerophosphate, 5 mM EGTA,
1 mM Na3VO4, and 1 mM dithiothreitol (Millipore Catalog #20-
108). The Mg/ATP cocktail contained 75 mM MgCl2, and 0.5 mM
cold ATP in assay dilution buffer (Millipore Catalog #20-113).
1. Schiffer, C. A. N. Eng. J. Med. 2007, 357, 258.
2. Capdeville, R.; Buchdunger, E.; Zimmermann, J.; Matter, A. Nat. Rev. Drug Disc.
2002, 1, 493.
3. Buchdunger, E.; Matter, A.; Druker, B. J. Biochim. Biophys. Acta 2001, 1551, M11.
4. Liu, Y.; Gray, N. S. Nat. Chem. Biol. 2006, 2, 358.
5. Nagar, B. et al Cancer Res. 2002, 62, 4236.
6. Regan, J. et al J. Med. Chem. 2003, 46, 4676.
7. Kroe, R. R. et al J. Med. Chem. 2003, 46, 4669.
8. Sherman, S. I. Endocrinol. Metab. Clin. North Am. 2008, 37, 511. xi.
9. Gupta-Abramson, V. et al J. Clin. Oncol. 2008, 26, 4714.
10. So, B. J.; Bekaii-Saab, T.; Bloomston, M. A.; Patel, T. J. Hematol. Oncol.
2008, 1, 18.
[gamma-32P) Adenosine-50-triphosphate (10
chased from Perkin–Elmer and used at 0.8 Ci/rxn. Stock 10 mM
lCi/lL) was pur-
l
drug solutions were made up in DMSO and from that were diluted
in a half-log manner in ddH2O.
11. Zender, L.; Kubicka, S. Onkologie 2008, 31, 550.
12. Wilhelm, S. M. et al Mol. Cancer Ther. 2008, 7, 3129.
13. Shim, J. H.; Park, J. W.; Choi, J. I.; Park, B. J.; Kim, C. M. J. Cancer Res. Clin. Oncol.
2009, 135, 617–625.
14. Garnett, M. J.; Marais, R. Cancer Cell 2004, 6, 313.
15. Ishimura, N. et al Cancer Lett. 2003, 199, 169.
16. Li, N.; Batt, D.; Warmuth, M. Curr. Opin. Investig. Drugs 2007, 8, 452.
17. Tuveson, D. A.; Weber, B. L.; Herlyn, M. Cancer Cell 2003, 4, 95.
18. Kumar, R. et al Clin. Cancer Res. 2003, 9, 3362.
19. Fukushima, T. et al Oncogene 2003, 22, 6455.
20. Ferrara, N.; Hillan, K. J.; Gerber, H. P.; Novotny, W. Nat. Rev. Drug Disc. 2004, 3,
391.
21. Ahmad, T.; Eisen, T. Clin. Cancer Res. 2004, 10, 6388S.
22. Eisen, T. et al Br. J. Cancer 2006, 95, 581.
23. Pargellis, C. et al Nat. Struct. Biol. 2002, 9, 268.
24. Schultz, R. M. Prog. Drug Res. 2003, 60, 59.
25. Lee, J. C. et al Immunopharmacology 2000, 47, 185.
26. Lee, J. C.; Kassis, S.; Kumar, S.; Badger, A.; Adams, J. L. Pharmacol. Ther. 1999, 82,
389.
27. Kumar, S.; Boehm, J.; Lee, J. C. Nat. Rev. Drug Disc. 2003, 2, 717.
28. Adams, J. L.; Badger, A. M.; Kumar, S.; Lee, J. C. Prog. Med. Chem. 2001,
38, 1.
29. Lee, M. R.; Dominguez, C. Curr. Med. Chem. 2005, 12, 2979.
30. Sullivan, J. E. et al Biochemistry 2005, 44, 16475.
31. Namboodiri, H. V. et al Biochemistry 2010, 49, 3611.
32. Fabian, M. A.; Biggs, W. H., IIIrd; Treiber, D. K.; Atteridge, C. E.; Azimioara, M. D.
et al. Nat Biotechnol 2005, 23, 329–336.
33. Karaman, M. W.; Herrgard, S.; Treiber, D. K.; Gallant, P.; Atteridge, C. E. et al.
Nat Biotechnol 2008, 26, 127–132.
34. Fedorov, O. et al Proc. Natl. Acad. Sci. 2007, 104, 20523.
35. Bankston, D. et al Org. Process Res. Dev. 2002, 6, 777.
36. Szakacs, Z. et al J. Med. Chem. 2005, 48, 249.
6.2. p38a Enzymatic assay
7.5 ng recombinant human full length N-terminal GST-tagged
p38 (Invitrogen, PV3304), Escherichia coli expression) was prein-
cubated at room temperature for 1 h with 1 L inhibitor and 4
assay dilution buffer. The kinase assay was initiated when 5 L of
a solution containing 200 ng recombinant human full length,
N-terminal His-tagged MAPKAP-K2 (sPV3316), 200 ATP
(0.8 Ci hot ATP), and 30 mM MgCl2 in assay dilution buffer was
added. The kinase reaction was allowed to continue at room tem-
perature for 25 min and was then quenched with 5 L 5X protein
denaturing buffer (LDS) solution. Protein was further denatured by
heating for 5 min at 70 °C. 10 L of each reaction was loaded into a
a
l
lL
l
l
M
l
l
l
15 well, 4–12% precast NuPage gel (invitrogen) and run at 200 V
and upon completion, the front which contained excess hot ATP
wascutfromthegeland discarded. Thegelwasthendriedand devel-
oped onto a phosphor screen which was scanned on a Storm 820
scanner and quantitated from optical densitometry using image
quant v5.0. A negative control which contained no active enzyme
was used as a negative control and a reaction without inhibitor
was used as the positive control. Final compound concentrations
were 100 lM, 31.6 lM, 10 lM, 3.16 lM, 1 lM, 316 nM, 100 nM,
31.6 nM, 10 nM, 3.16 nM, 1 nM, 316 pM, and 100 pM.
6.3. B-Raf enzymatic assay
37. Lutz, F.; Tietze, F. M. Eur. J. Org. Chem. 2006, 2006, 2314.
38. Baldwin, I. et al Bioorg. Med. Chem. Lett. 2008, 18, 5285.
39. Baranska, M. et al J. Appl. Genet. 2008, 49, 201.
The B-Raf biological assay used the same protocol as p38 except
it utilized recombinant Human Full-Length GST-tagged B-Raf
(invitrogen, pv3848) and recombinant human full length, N-termi-
nal His-tagged MEK1 (invitrogen, pv3093).