2-Amino-7-substituted benzoxazole analogs as potent RSK2 inhibitors

Article abstract of DOI:10.1016/j.bmcl.2014.01.058

2-Amino-7-substituted benzoxazole analogs were identified by HTS as inhibitors of RSK2. Molecular modeling and medicinal chemistry techniques were employed to explore the SAR for this series with a focus of improving in vitro and target modulation potency and physicochemical properties.

Full text of DOI:10.1016/j.bmcl.2014.01.058

Bioorganic & Medicinal Chemistry Letters 24 (2014) 1592–1596
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
2-Amino-7-substituted benzoxazole analogs as potent RSK2
inhibitors
Abran Costales, Michelle Mathur, Savithri Ramurthy, Jiong Lan, Sharadha Subramanian, Rama Jain,
Gordana Atallah, Lina Setti, Mika Lindvall, Brent A. Appleton, Elizabeth Ornelas, Paul Feucht, Bob Warne,
⇑
Laura Doyle, Stephen E. Basham, Ida Aronchik, Anne B. Jefferson, Cynthia M. Shafer
Global Discovery Chemistry/Oncology & Exploratory Chemistry, Novartis Institutes for Biomedical Research, 4560 Horton Street, Emeryville, CA 94608, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
2-Amino-7-substituted benzoxazole analogs were identified by HTS as inhibitors of RSK2. Molecular
modeling and medicinal chemistry techniques were employed to explore the SAR for this series with a
focus of improving in vitro and target modulation potency and physicochemical properties.
Ó 2014 Elsevier Ltd. All rights reserved.
Received 5 December 2013
Revised 14 January 2014
Accepted 21 January 2014
Available online 30 January 2014
Keywords:
RSK
Kinase
Heterocycle
Benzoxazole
HTS
The p90 Ribosomal S6 kinases (RSK) are a family of 4 serine/
threonine kinases widely expressed across tissues and activated
in response to numerous hormones and growth factors.1 The RSK
isoforms have a unique structure containing two nonidentical ki-
nase domains (N-terminal and C-terminal) that are separated by
a linker region. A complex cascade of phosphorylation events is re-
quired for RSK activation. The current favored model for RSK acti-
vation entails ERK activating RSK’s C-terminal kinase domain
which then provides a phosphorylation-based docking site for
PDK1 to bind and activate RSK’s N-terminal kinase domain. The
N-terminal kinase then functions to phosphorylate numerous nu-
clear and cytoplasmic proteins that account for the ascribed cellu-
lar roles of RSK including cell survival, proliferation and motility.2
Its unusual activation scheme positions RSK to integrate signal-
ing inputs from both the MAPK and PDK1 signaling pathways. Both
of these pathways are relevant to oncology, as mutational activa-
tion of K-Ras and B-Raf are among the best validated oncogenic
drivers of human cancers. Tumor cell lines with these mutations
also have strong phosphorylation of RSK activation sites, and re-
cent publications point to mutationally activated FGFR as an addi-
tional cancer-relevant kinase that can phosphorylate RSK and
sensitize it to activation via MAPK pathway signaling.3
Despite the potential interest of RSK as a target for oncology
therapeutics, there are relatively few chemical tools to explore
the importance of RSK in cancer.4
High-throughput screening (HTS) yielded the 2-arylamino-7-
phenylbenzoxazole compound 1 (Fig. 1) as an N-terminal inhibitor
of the RSK1–4 isoforms. This compound was derived from a series
developed by the JAK2 project at Novartis.5 Although compound 1
was only modestly potent against RSK2, it showed excellent selec-
tivity against a panel of over 80 kinases suggesting that it would be
a good starting point to improve enzymatic affinity while main-
taining selectivity. Further, conformational analysis of 1 in the
gas phase and in water suggested a torsion angle of 48–54 degrees
around the biaryl bond and a high energy barrier to achieve planar-
ity of these two rings. Preference for this non-planar conformation
was hypothesized to contribute to the kinase selectivity observed.
The concern regarding this scaffold was poor solubility. Thus, strat-
egies involving reducing the aromaticity of the scaffold, lowering
the cLogP or incorporating a basic amine were developed.
However, prior to addressing the physicochemical attributes of
the scaffold, an understanding of the basic SAR of the scaffold against
RSK2 was undertaken. The SAR of the left hand portion of the mole-
cule was first evaluated (Table 1). Removing the aryl ring to give the
2-aminobenzoxazole 2 resulted in a 7-fold loss of potency while re-
moval of the meta-methoxy moiety from the phenyl ring (3) resulted
in a 2-fold loss in affinity. Interestingly, the saturation of the phenyl
ring to give the cyclohexyl ring (4) was equipotent with 1. The
⇑
Corresponding author. Tel.: +1 510 923 8086.
E-mail address: cynthia.shafer@novartis.com (C.M. Shafer).
http://dx.doi.org/10.1016/j.bmcl.2014.01.058
0960-894X/Ó 2014 Elsevier Ltd. All rights reserved.

A. Costales et al. / Bioorg. Med. Chem. Lett. 24 (2014) 1592–1596
1593
had 20-fold better affinity than 1. Compound 8 was also the first
example in this series that had single digit micromolar target mod-
ulation against the downstream marker of interest, p-YB1. How-
ever, the selectivity profile for 8 was marginally worse than for 1.
We hypothesized that perhaps the RSK2 biochemical affinity could
be retained while undesired kinase activities could be eliminated
by determining if specific methoxy groups were required for affin-
ity while others contributed to the loss of selectivity. Unfortu-
nately, none of the examples made (9–11) retained the affinity of
8 although the 3,4-dimethoxyaniline analog (11) was only 3-fold
less potent than 8. This indicated that all three methoxys were re-
quired for potent inhibition of RSK2.
It was thought that hydrogen bonding groups in the meta posi-
tion could interact with neighboring polar residues directly or
through water molecules. To test this hypothesis, the sulfonamide
(12) was prepared which presented both a hydrogen bond acceptor
and a hydrogen bond donor to the protein for a possible hydrogen
bonding opportunity with Asp154. However, 12 was equipotent to
the original HTS hit 1. A hydrogen bond donor along with a
para-methoxy group was incorporated into 13 to attempt to hydrogen
bond to Asp154 and allow for protein flexibility; however, 13 was
Figure 1. RSK 2-arylamino-7-arylbenzoxazole HTS hit.
tetrahydropyran (5), however, was 5-fold less potent than 4 suggest-
ing that while hydrophobic bulk is necessary in this area of the pock-
et for affinity, a hydrogen bond acceptor at this position is not
accepted. In an attempt to lower the cLogP of the scaffold through
incorporation of additional heteroatoms, the pyridine 6 was synthe-
sized and found to be equipotent to 1. Unfortunately, the addition of
the nitrogen altered the cLogP very little (5.4 for 1 vs 5.1 for 6), and
no improvement in the solubility was realized.
Continuing to explore the SAR around the phenyl ring, the meta-
ethoxy analog (7) was synthesized which had little effect on affin-
ity. Incorporating a 3,4,5-trimethoxyphenyl group gave 8 which
Table 1
Structure activity relationship of the benzoxazole aniline for RSK2
N
HN
R1
O
F
O
Compound
R1
RSK2 IC50
0.41
(
lM)
p-YB1 TM (
lM)
Compound
R1
RSK2 IC50
(lM)
p-YB1 TM (
lM)
O
1
15.4
>20
>20
8
0.02
2.6
O
O
O
2
3
H
2.8
1.0
9
1.4
>20
O
O
10
0.21
17.2
O
O
4
5
6
0.62
3.2
>20
>20
>20
11
12
13
0.07
0.34
0.22
6.5
O
O
>20
>20
O
O
S O
NH2
N
0.62
O
O
NH2
O
7
0.73
>20
14
0.01
6.3
O
HN

1594
A. Costales et al. / Bioorg. Med. Chem. Lett. 24 (2014) 1592–1596
Figure 2. Crystal structure (PDB 4NW5) of 8 with Rsk2 NTKD.
Figure 3. Crystal structure of 27 (PDB 4NW6) with Rsk2 NTKD.
similarly potent to 1 suggesting that the required trajectory or
length was not attained. Analysis of the crystal structure of 8
(Fig. 2) suggested that a basic amine might be able to form a hydro-
gen bond with Asp154 as well as improve physicochemical proper-
ties. However, the crystal structure of 14 indicates that no
additional hydrogen bonds are made between the pyrrolidine
and the protein backbone. Additional van der Waals contacts
may explain the potency boost. Compound 14 exhibited the best
affinity and solubility observed thus far; however kinase selectivity
compared to 1 was significantly worse and thus basic amines from
this trajectory were not further pursued.
With an understanding of the SAR of the arylamino portion of
the scaffold mapped out, we turned our attention to the evaluation
of the C-7 aryl substituent (Table 2). Removal of the fluoro (15) did
not affect the biochemical affinity compared to 8. Conformational
analysis by quantum mechanics suggested that the two conforma-
tions where methoxy is either syn or anti to benzoxazole oxygen
were equally energetically accessible with or without fluorine.
Table 2
Structure activity relationship of the C7 of the benzoxazole for RSK2
N
HN
O
O
R2
O
O
Compound
R2
RSK2 IC50
0.02
(lM)
p-YB1 TM (
l
M)
Compound
R2
RSK2 IC50
0.04
(lM)
p-YB1 TM (
lM)
F
O
O
8
2.6
21
3.7
F
O
O
O
15
0.02
5.7
22
0.04
9.7
F
O
16
17
0.22
4.0
18
23
24
1.1
9.3
14.5
>20
NH
N
CF3
O
>20
F
F
N
18
19
0.49
0.97
10.8
>20
25
26
0.56
0.05
19.8
5.7
HN
H
N
O
N
O
O
N
20
1.2
>20
27
0.004
0.8
N
H

A. Costales et al. / Bioorg. Med. Chem. Lett. 24 (2014) 1592–1596
1595
N
O
N
N
(a)
HN
f
HN
O
H2N
HO
a
NO2
HO
b,c
d,e
S
O
F
O
O
O
Br
O
Br
Br
Br
O
O
O
28
29
30
31
8
a Reagents and conditions:(a) Zn, NH4Cl, MeOH, 97%;(b) Potassium methylxanthate, EtOH, 90 oC, 80%; (c) MeI, K2CO3, DMF, 75%; (d) m-CPBA, DCM,
94%; (e) 3,4,5-Trimethoxyaniline, NMP, 120 oC, 40%; (f) 2-fluoro-6-methoxyphenyl boronic acid, Na2CO3, DME, PdCl2(dppf), 120 oC, 19%.
N
O
N
O
N
O
(b)
HN
O
HN
O
HN
O
b
a
O
O
O
B
Br
N
O
O
O
O
O
N
H
31
32
27
a Reagentsand conditions:(a)bis (pincacolato)diboron, 1,4-Dioxane,KOAc,Pd(dba)2, Cy3P,120 oC, 100%; (b) 4-bromo-1H-benzo[d]imidazole,
Na2CO3, DME, PdCl2 (dppf), 120o C, 10%.
Scheme 1. Synthetic methods for 7-substituted-benzo[d]oxazol-2-amine series.
Removal of methoxy (16) gave a 10-fold reduction in affinity. Alter-
ation of the methoxy to a trifluoromethoxy (17) led to a dramatic
loss in affinity (200-fold) which is consistent with the bound struc-
ture of 8 (Fig. 2) where trifluoromethyl in 17 weakens the H-bond
to Thr210 and causes a steric and electrostatic clash between the
trifluoromethyl group and the protein. Extension of the methoxy
to an ethoxy (18) or isopropoxy (19) gave 25- and 50-fold potency
losses, respectively. Replacement of the 2-fluoro group with a
methoxy (20) led to a significant loss in potency. Moving the fluoro
group to the meta position (21) resulted in similar potency to 8.
Incorporation of a 4-methyl group (22) retained affinity while
moving the methyl to the meta-position (23), led to a 28-fold po-
tency loss. Docking 23 into the crystal structure of 8 (Fig. 2) sug-
gests a steric clash between the methyl and the P loop.
can form a hydrophobic contact with Leu74 of the hydrophobic
lower hinge channel of kinases, leading to an increase in potency.
The 7-substituted-benzo[d]oxazol-2 amine analogs described
herein were synthesized following the procedures of Gerspacher
et al.5a (Scheme 1a) or by a slight variant (Scheme 1b). Compound
8 (Scheme 1a) was synthesized starting from commercially avail-
able 2-bromo-6-nitrophenol (28) which underwent selective nitro
reduction followed by condensation with potassium ethylxanthate
and subsequent methylation to afford 7-bromo-2-(methylthio)-
benzo[d]oxazole (30). The remaining sequence involved activation
of the sulfur leaving group of 30 with m-CPBA followed by
displacement with commercially available 3,4,5-trimethoxyanline
to give 31. Finally, Suzuki coupling with commercially available
2-fluoro-6-methoxyphenyl boronic acid afforded 8. Alternatively,
compound 31 (Scheme 1b) can be borolated followed by Suzuki
coupling with 4-bromo-1H-benzo[d]imidazole to afford 27.
The co-crystal structure of compound 8 (Fig. 2) shows a
hydrogen bond between the oxygen of the methoxy and Thr210.
A
bi-dentate hydrogen bond is also revealed between the
In conclusion, from a 0.4 lM HTS hit (1), the highly potent and
benzoxazole core and Leu150 of the hinge.
selective 27 was designed utilizing structure based drug design
(SBDD) and synthesized. With compound 27, a 100-fold potency
improvement was realized concurrent with retaining a 100+-fold
selectivity profile against a panel of diverse kinases. In addition,
27 exhibited sub-micromolar target modulation which correlated
well with the inhibition of cell proliferation on low bind plates.
However, an optimal combination of physicochemical properties,
selectivity and in vitro potency was not obtainable during the
course of the program, and the scaffold was not pursued further.
Further modifications could be envisioned which combine the po-
tency-enhancing properties of the benzimidazole with a left hand
piece which would disrupt crystal packing.
With this knowledge, we attempted to reach threonine 210 with
other moieties. The 2-(4-pyrazolo)phenyl analog (24) was not ac-
tive, consistent with steric clash upon docking. Replacement of
the phenyl ring with a 2,5-dimethylimidazole (25) indicated that
the imidazole nitrogen was in the correct vicinity to make the con-
tact with the threonine but is not optimal. And finally the 5,6-fused
heterocycles (26–27) were synthesized. While the 4-benzimidazole
(27) was the most potent analog synthesized in this series, 26 was
highly selective. The selectivity of 27 was also quite good. In a panel
of 35 kinases, 33 kinases had IC50s >10
lM with Aurora A and JAK2
being inhibited with IC50s of 0.6 M and 0.46
l
lM respectively.
Compound 27 exhibited submicromolar inhibition of target modu-
lation and cellular proliferation in of the human cancer cell line
Acknowledgment
MDA-MB-231 on low bind plates (EC50 0.58 lM). However the sol-
ubility was not improved compared to 1. Finally, the co-crystal
structure of 27 was also determined (Fig. 3). In comparison to 8,
the benzimidazole replaces the hydrogen bond to the side-chain
of Thr210. Qualitative energetic analysis suggests that the benz-
imidazole acts as a hydrogen bond acceptor and Thr210 hydroxyl
group as hydrogen bond donor, as in the crystal structure of 8.
In both co-crystal structures, the para methoxy oxygen of the
trimethoxy aniline is forced out of plane by the 3,5-dimethoxy
groups. In such conformation, the methyl group of 4-methoxy
The authors would like to thank Dahzi Tang for his analytical
support.
Supplementary data
Supplementary data (Experimental details for the synthesis
and characterization of select compounds, procedures for the
biochemical and cellular assay and additional information on the
computational methods utilized are included in the supporting

1596
A. Costales et al. / Bioorg. Med. Chem. Lett. 24 (2014) 1592–1596
3. (a) Kang, S.; Dong, S.; Gu, T.-L.; Guo, A.; Cohen, M. S.; Lonial, S.; Khoury, H. J.;
Fabbro, D.; Gilliland, D. G.; Bergsagel, P. L.; Taunton, J.; Polakiewicz, R. D.; Chen,
J. Cancer Cell 2007, 12, 201; (b) Kang, S.; Elf, S.; Dong, S.; Hitosugi, T.; Lythgoe, K.;
Guo, A.; Ruan, H.; Lonial, S.; Khoury, H. J.; Williams, I. R.; Lee, B. H.; Roesel, J. L.;
Karsenty, G.; Hanauer, A.; Taunton, J.; Boggon, T. J.; Gu, T.-L.; Chen, J. Mol. Cell.
Biol. 2009, 29, 2105; (c) Xian, W.; Pappas, L.; Pandya, D.; Selfors, L. M.; Derksen,
P. W.; de Bruin, M.; Gray, N. S.; Jonkers, J.; Rosen, J. M.; Brugge, J. S. Cancer Res.
2009, 69, 2244.
4. (a) Nguyen, T. L. Anti-Cancer Agents Med. Chem. 2008, 8(7), 710; (b) Andreani, A.;
Granaiola, M.; Leoni, A.; Locatelli, A.; Morigi, R.; Rambaldi, M.; Varoli, L.;
Lannigan, D.; Smith, J.; Scudiero, D.; Kondapaka, S.; Shoemaker, R. H. Eur. J. Med.
Chem. 2011, 46, 4311; (c) Boyer, S. J.; Burke, J.; Guo, X.; Kirrane, T. M.; Snow, R. J.;
Zhang, Y.; Sarko, C.; Soleymanzadeh, L.; Swinamer, A.; Westbrook, J.; DiCapua,
F.; Padyana, A.; Cogan, D.; Gao, A.; Xiong, Z.; Madwed, J. B.; Kashem, M.; Kugler,
S.; O’Neill, M. M. Bioorg. Med. Chem. Lett. 2012, 22, 733.
information.) associated with this article can be found, in the on-
line version, at http://dx.doi.org/10.1016/j.bmcl.2014.01.058.
References and notes
1. (a) Anjum, R.; Blenis, J. Nat. Rev. Mol. Cell Biol. 2008, 9, 747; (b) Carriere, A.; Ray,
H.; Blenis, J.; Roux, P. P. Front. Biosci. 2008, 13, 4258.
2. (a) Chen, R. H.; Abate, C.; Blenis, J. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 10952; (b)
Shimamura, A.; Ballif, B. A.; Richards, S. A.; Blenis, J. Curr. Biol. 2000, 10, 127; (c)
Bonni, A.; Brunet, A.; West, A. E.; Datta, S. R.; Takasu, M. A.; Greenberg, M. E.
Science (Washington, DC) 1999, 286, 1358; (d) Roux, P. P.; Shahbazian, D.; Vu, H.;
Holz, M. K.; Cohen, M. S.; Taunton, J.; Sonenberg, N.; Blenis, J. J. Biol. Chem. 2007,
282, 14056; (e) Carriere, A.; Cargnello, M.; Julien, L.-A.; Gao, H.; Bonneil, E.;
Thibault, P.; Roux, P. P. Curr. Biol. 2008, 18, 1269; (f) Doehn, U.; Hauge, C.; Frank,
S. R.; Jensen, C. J.; Duda, K.; Nielsen, J. V.; Cohen, M. S.; Johansen, J. V.; Winther, B.
R.; Lund, L. R.; Winther, O.; Taunton, J.; Hansen, S. H.; Frodin, M. Mol. Cell 2009,
35, 511; (g) Stratford, A. L.; Fry, C. J.; Desilets, C.; Davies, A. H.; Cho, Y. Y.; Li, Y.;
Dong, Z.; Berquin, I. M.; Roux, P. P.; Dunn, S. E. Breast Cancer Res. 2008, 10, R99.
5. (a) Gerspacher, M.; Furet, P.; Pissot-Soldermann, C.; Gaul, C.; Holzer, P.;
Vangrevelinghe, E.; Lang, M.; Erdmann, D.; Radimerski, T.; Regnier, C. H.;
Chene, P.; De Pover, A.; Hofmann, F.; Baffert, F.; Buhl, T.; Aichholz, R.; Blasco, F.;
Endres, R.; Trappe, J.; Drueckes, P. Bioorg. Med. Chem. Lett. 2010, 20, 1724; (b)
Gerspacher, M.; Furet, P.; Vangrevelinghe, E. WO 2,008,031,594, 2008.