4
S. L. Raeppel et al. / Bioorg. Med. Chem. Lett. xxx (2015) xxx–xxx
O
O
1.
O
O
O
O
O
DCC, DMAP, DCM
C to rt
O
O
O
N
2
Me N
O
BocHN
o
0
OMe
NMe
BocHN
BocHN
OH
OMe
DCM
reflux
2
. MeOH, reflux
14
2
13
NHNH
2
N
N
N
3
1. Et N, DMF
MeO
2
C
HCl/dioxane
o
MeO
2
C
130 C
N
AcOH
MeOH
reflux
EtOAc/EtOH
o
2. NaOMe
MeOH
N
15
0
C to rt
Cl-
HN
16
BocHN
+H
N
0 C to rt
o
3
O
1
7
F
Br
Compound 17
Pd (dba)
2 3
Xantphos
F
Br
Cl
N
N
N
HO
F
N
N
S
O
N
O
K
Ph
2
CO
3
N
S
Cs CO3
O
2
2
O
dioxane
1
8
o
N
19
N
S
210 C
N
100 C
o
20
N
N
Scheme 3. Synthesis of compound 20.
Compound 21 featuring a [6,5]-bicyclic head group was pre-
pared in one step by a palladium catalyzed N-arylation coupling
between intermediates 19 and 1-phenyl-1,5,6,7-tetrahydro-4H-
pyrazolo[4,3-c]pyridin-4-one16 using conditions similar to the
preparation of compound 20 (Scheme 3, step 7).
Our design was guided by molecular docking studies using an
homology model of RON kinase domain built from the X-ray
crystal structure of c-Met kinase domain (PDB: 3U6I)19 which
shares a high sequence homology with RON. The docking of
20
LCRF-0004 using the Fitted program revealed a class II extended
2
1
conformation with the head group (Part C) deeply engaged in the
hydrophobic back pocket of the enzyme interacting with the back-
bone NH of Asp1226. A probable intramolecular hydrogen bond
between one fluorine atom from the trifluoromethyl group and
the NH-acidic carboxamide seems to occur in order to rigidify
the head group in the desired bioactive conformation. The
thieno[3,2-b]pyridine scaffold (Part B) interacts with the hinge
binding domain (NH of Met1164) and 1-methyl-1H-imidazole sub-
stituent (Part A) is positioned in the solvent-exposed area (Fig. 3).
The docking of LCRF-0004 in our homology model of the active site
of RON showed a predicted docking score of À24.6 versus À18.6 for
compound 1. Predicted docking score for compounds 12 and 21
were À22.1 and À22.9, respectively.
Compound 4 was synthesized as the first prototype to verify our
hypothesis of a [7,5]-bicyclic structure as a constrained analog of
LCRF-0004. Although less active against RON and c-Met enzymes,
3
6 and 17 times, respectively, when compared to LCRF-0004, this
new prototype is still interesting for innovation aspects.
Fortunately, the replacement of the amido group in compound 4
by a double bond as in compound 12 was beneficial in restoring
the inhibition of the RON enzymatic activity at the expense of
c-Met. The saturation of the double bond as in [7,5]-bicyclic com-
pound 20 improves slightly the activity for RON, but c-Met as well.
Likewise, the presence of a more compact [6,5]-bicyclic system 21
seems to have a better affinity for both RON and c-Met kinase
domains by slightly shifting the orientation of both the lactam
carbonyl and the phenyl substituent on the pyrazole ring. Thus,
in order to maintain good selectivity for RON over c-Met, it is
preferable to further explore the [7,5]-bicyclic system by adding
more decorations on this part of the molecule that would allow
us to increase the inhibitory activity for RON.
Figure 3. Predicted binding mode in an homology model of RON kinase domain
from docking. (a) LCFR-0004; (b) compound 12; (c) compound 21.
In conclusion, we have designed and synthesized new potent
inhibitors of RON based on LCRF-0004, as mimetic of its potential
bioactive conformation. Our results allowed us to support this
hypothesis and study the effect of diverse fused bicyclic head
2
2
groups on the RON kinase inhibitory activity and selectivity over
c-Met enzyme. Based on these results, we were able to design and
synthesize a third series of potent RON inhibitors that will be