Design and synthesis of 7H-pyrrolo[2,3-d]pyrimidines as focal adhesion kinase inhibitors. Part 2

Article abstract of DOI:10.1016/j.bmcl.2006.02.032

A series of 2-amino-9-aryl-7H-pyrrolo[2,3-d]pyrimidines were designed and synthesized as focal adhesion kinase (FAK) inhibitors using molecular modeling in conjunction with a co-crystal structure. Chemistry was developed to introduce functionality onto th

Full text of DOI:10.1016/j.bmcl.2006.02.032

Bioorganic & Medicinal Chemistry Letters 16 (2006) 2689–2692
Design and synthesis of 7H-pyrrolo[2,3-d]pyrimidines as
focal adhesion kinase inhibitors. Part 2
Ha-Soon Choi,a Zhicheng Wang,a Wendy Richmond,a Xiaohui He,a Kunyong Yang,a
Tao Jiang,a Donald Karanewsky,a Xiang-ju Gu,a Vicki Zhou,a Yi Liu,a Jianwei Che,a
Christian C. Lee,a Jeremy Caldwell,a Takanori Kanazawa,b Ichiro Umemura,b
Naoko Matsuura,b Osamu Ohmori,b Toshiyuki Honda,b Nathanael Graya and Yun Hea,*
aGenomics Institute of the Novartis Research Foundation (GNF), 10715 John Jay Hopkins Drive, San Diego, CA 920121, USA
bNovartis Institute for BioMedical Research Tsukuba, Okubo 8, Tsukuba-shi, Ibaraki 300-2611, Japan
Received 16 January 2006; revised 6 February 2006; accepted 7 February 2006
Available online 9 March 2006
Abstract—A series of 2-amino-9-aryl-7H-pyrrolo[2,3-d]pyrimidines were designed and synthesized as focal adhesion kinase (FAK)
inhibitors using molecular modeling in conjunction with a co-crystal structure. Chemistry was developed to introduce functionality
onto the 9-aryl ring, which resulted in the identification of potent FAK inhibitors. In particular, compound 32 possessed single-digit
nanomolar IC50 and represents one of the most potent FAK inhibitors discovered to date.
Ó 2006 Elsevier Ltd. All rights reserved.
Focal adhesion kinase (FAK) is a protein-tyrosine
kinase that is found at the sites of cellular contact and
is phosphorylated in response to cell attachment.1–3
FAK plays an important role in cellular movement
7
N
8
9
N
HN
2
N
2'
and survival pathways. FAK is a potential target
for the treatment of both primary cancers and the
prevention of tumor metastasis.4–7 In the previous
communication, we reported the synthesis of a series
of 7H-pyrrolo[2,3-d]pyrimidines as inhibitors of FAK.8
Among these, compound 1 exhibited submicromolar
inhibitory activity against FAK (Fig. 1). A co-crystal
structure of compound 1 with FAK was also successful-
ly obtained.9 Molecular modeling based on the co-crys-
tal structure suggested that the inhibitory activity of this
series of compounds could potentially be improved by
introducing additional interactions between the K454,
D564, R550 or E502 amino acid residues of the kinase
and the inhibitor.
N
3'
MeO
OMe
4'
OMe
1
Figure 1. Structure of pyrrolopyrimidine 1 and its complex with FAK.
Compound 7 was synthesized in six steps starting
from commercially available 3-bromo-2-oxo-propionic
acid 2.10 After a two-step conversion of 2 to methyl
enol ether ester 3,11 the corresponding vinyl stannane
4 was prepared by reaction with tributylstannyl cop-
per. Palladium-catalyzed coupling of 4 with bromo-
pyrimidine 5 gave the corresponding vinyl ether in
good yield.8,12 Treatment of the resulting vinyl ether
with hydrochloric acid furnished the pyrrolopyrimidi-
nyl ester 6, which upon saponification, led to the
desired carboxylic acid 7. Contrary to our expectation,
the appendage of the carboxy group at position 8
resulted in a complete loss of inhibitory activity
(Table 1).
Our initial attempt focused on forming a salt bridge
or hydrogen bond to K454 by introducing a carboxylate
to the 8-position of pyrrolopyrimidine (7, Scheme 1).
Keywords: 7H-Pyrrolo[2,3-d]pyrimidines; Focal adhesion kinase
(FAK); Molecular modeling; SAR; Kinase inhibitor.
*
Corresponding author. Tel.: +1 858 332 4706; fax: +1 858 332
4351; e-mail: yhe@gnf.org
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.02.032

2690
H.-S. Choi et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2689–2692
O
OMe
CO2Me
Br
OMe
c
a,b
N
Br
Br
NO2
NH2
Bu3Sn
CO2Me
CO2H
a,b
SO2Cl
SO2NHPr
c
Cl
N
NH
2
3
4
SO2NHPr
d,e
8
9
10
Br
N
N
N
CO2H
CO2Me
N
N
HN
N
HN
N
HN
N
NH
N
d
f
N
N
N
OEt
NH
N
N
MeO
MeO
MeO
OMe
OMe
N
OMe
SO2NHnPr
HN
N
N
OMe
OMe
f
e
OMe
Cl
N
Cl
N
PrHNO2S
7
6
5
PrHNO2S
MeO
OMe
Scheme 1. Synthesis of
6 and 7. Reagents and conditions: (a)
OMe
13
12
11
(MeO)3CH (6.0 equiv), MeOH, H2SO4 (cat), reflux, 4 h; (b) p-TSOH
(cat), 210 °C, 3 h, 68% for two steps; (c) CuSnBu3, THF, ꢀ78 °C to
25 °C, 3 h, 47%; (d) 5 (1.0 equiv), Pd(PPh3)4 (cat), toluene, reflux, 4 h,
81%; (e) HCl (3.0 N), i-PrOH, reflux, 4 h, 80%; (f) NaOH (2.0 N),
MeOH, 25 °C, 5 h, 86%.
Scheme 2. Synthesis of 13. Reagents and conditions: (a) n-PrNH2
(1.5 equiv), Et3N (2 equiv), 23 °C, 5 h; (b) 10% Pd/C (cat), H2, EtOH,
25 °C, 16 h; (c) 5-bromo-2,4-dichloropyrimidine (1.0 equiv), 80%
AcOH, 50 °C, 16 h, 52% for three steps; (d) EtOCHCHSnBu3
(1.2 equiv), Pd(PPh3)4, (cat) toluene, 110 °C, 1 h, 83%; (e) HCl
(3.0 N), i-PrOH, reflux, 4 h, 90%; (f) 3,4,5-trimethoxyaniline
(1.5 equiv), t-BuOK, 1,4-dioxane, 70 °C, 8 h, 45%.
Table 1. Enzymatic IC50 against FAK for inhibitors 1–328
Cpd
IC50 (lM)
Cpd
IC50 (lM)
Cpd
IC50 (lM)
1
0.212
>10
>10
18b
18c
18d
18e
18f
18g
18h
18i
19
1.53
5.12
23l
0.983
0.434
0.882
1.81
In an effort to understand why our designed inhibitors 7
and 13 were significantly less active than our lead com-
pound 1, we used molecular mechanics to study the
conformational preferences of our inhibitors. These
studies suggested that the diminished FAK inhibitory
activity might result from the compounds being
predisposed to a conformation less favorable to binding.
Specifically, the mechanics calculations suggest that the
9-pyridyl ring in 1 and the 9-aryl ring in 13 have
significantly different orientations in the ground state
(Fig. 2). The aryl ring in 13 is nearly perpendicular to
the pyrrolopyrimidine ring in contrast to the nearly
planar orientation of the pyridine ring in 1. Similar
results were obtained for compound 7. Since these
conformational changes are likely caused by the steric
hindrance created between the pyrrole moiety and the
9-aryl moiety, we decided to pursue substituents that
would minimize these interactions.
6
23m
23n
24a
24b
24c
24d
24e
24f
24g
24h
24i
7
13
7.32
3.50
2.81
0.831
1.51
3.23
16a
16b
16c
16d
16e
16f
16g
16h
16i
17a
17b
17c
17d
17e
17f
17g
17h
17i
17j
17k
17l
18a
0.421
0.117
5.80
>10
4.05
6.50
2.84
>10
0.727
>10
0.727
1.33
>10
0.115
0.232
0.119
>10
20
21
0.504
0.314
2.82
0.635
0.318
0.331
0.345
0.063
0.236
0.339
7.55
22a
22b
22c
22d
23a
23b
23c
23d
23e
23f
23g
23h
23i
23j
23k
0.342
>10
>10
25a
25b
25c
25d
25e
28a
28b
28c
28d
28e
28f
28g
28h
32
0.221
0.235
0.126
0.063
0.093
1.507
0.038
0.037
0.035
2.71
>10
0.262
0.365
0.322
5.06
>10
0.182
0.634
0.397
0.278
0.242
0.544
0.004
4.11
0.029
0.038
0.073
0.298
0.054
A library of analogs with various substitutions on the
9-phenyl moiety (20, 30 and 40) was synthesized using
copper-mediated coupling of phenyl bromides with pyr-
rolopyrimidine intermediate 15 (Scheme 3). Consistent
with what was observed for 13, most of the analogs with
ortho substitutions (16a–16h) exhibited weak inhibitory
activity. Among these analogs, only the ester (16h) and
nitrile (16g) analogs showed 0.314 lM and 0.504 lM
IC50, respectively. In contrast, most of the 30-substituted
analogs (17a–17l) demonstrated comparable or im-
proved inhibitory activity compared to 1 (0.212 lM).
In particular, the carboxy analogs, 17g–17i, represent
significantly improved potency, suggesting that these
pyrrolopyrimidines might have gained additional inter-
actions with the enzyme via these substitutions. The
40-substituted analogs (18a–18i) did not exhibit any
improved activity. With this information in hand, we
prepared an additional library of 30-substituted analogs
with various functional groups that have the potential to
form interactions with the enzyme. The nitrile analogs
(17c and 17d) were thus converted into the correspond-
ing tetrazoles (19–21) in good yields by treatment with
>10
>10
0.507
We next tried to introduce a contact to either K454 or
E502 by incorporating a sulfonamide appendage on
the 9-phenyl moiety (13, Scheme 2). Compound 13
was synthesized in six steps from commercially available
sulfonyl chloride 8. Selective displacement of 5-bromo-
2,4-dichloropyrimidine with aniline 9 gave the corre-
sponding anilinyl pyrimidine 10 in good yield. Stille
coupling of the resulting 5-bromopyrimidine led to vinyl
ether 11, which was cyclized to pyrrolopyrimidine 12 in
excellent yield using hydrochloric acid. Displacement of
the chlorine in 12 with 3,4,5-trimethoxyaniline gave the
desired sulfonamide analog 13 in moderate yield. Again,
to our disappointment, sulfonamide 13 displayed signif-
icantly reduced FAK inhibitory activity relative to 1
(Table 1).

H.-S. Choi et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2689–2692
2691
Figure 2. Ground state conformations of 1 and 13.13
7
N
N
N
N
8
N
N
9
HN
N
HN
N
N
N
a,b
2'
2'
4'
HN
N
HN
N
OEt
H
N
2'
b
a
3'
3'
CONHR
CO2H
3'
( )n
4' ( )n
Cl
N
NH2
MeO
OMe
MeO
OMe
4'
MeO
OMe
MeO
OMe
R
OMe
OMe
OMe
OMe
16i: 2'-CO2H
17g: 3'-CO2H
17h: 3'-CH2CO2H
17i: 3'-(CH2)2CO2H
22: 2'-CONHR
23: 3'-CONHR
24: 3'-CH2CONHR
25: 3'-(CH2)2COHR
14
15
16-18
16a:
16b:
16c:
16d:
16e:
16f:
16g:
16h:
16i:
17b: R=3'-CF3
17c: R=3'-CN
R=2'-Me
R=2'-Et
17l: R=3'-CH2NH2
18a: R=4'-Me
17d: R=3'-CH2CN
17e: R=3'-(CH2)2CN
17f: R=3'-CO2Me
17g: R=3'-CO2H
R=2'-NO2
R=2'-NH2
R=2'-CH2OH
R=2'-NHAc
R=2'-CN
R=2'-CO2Me
R=2'-CO2H
R=3'-Me
18b: R=4'-Et
22a: R=H
23e:R=iBu
23f: R=cPent
23m:R=(CH2)2OH 24g: R=(CH2)3NMe2
23n: R=(CH2)3OH 24h: R=CH2CO2Me
18c: R=4'-nBu
18d: R=4'-tBu
18e: R=4'-SMe
18f: R=4'-CF3
18g: R=4'-NO2
18h: R=4'-CO2Me
18i: R=4'-CO2H
22b: R=Me
22c: R=(CH2)2NH2 23g:R=(CH2)2NH2
22d: R=(CH2)3NMe2 23h:R=(CH2)2NMe2
23a: R=H
23b: R=Me
23c: R=Et
23d: R=cPr
24a: R=Me
24b: R=Et
24i: R=(CH2)2CO2Me2
25a: R=Me
25b: R=Et
25c: R=iBu
17h: R=3'-CH2CO2H
23i: R=(CH2)2N(CH2)4 24c: R=cPr
17i: R=3'-(CH2)2CO2H
17j: R=3'-SO2NHMe
17k: R=3'-NH2
23j: R=(CH2)2NHAc
23k:R=(CH2)3NMe2
23l: R=(CH2)4NMe2
24d: R=cPent
24e: R=(CH2)2NH2 25d: R=(CH2)2NMe2
24f: R=(CH2)3NH2 25e: R=CH2CO2Me
17a:
Scheme 3. Synthesis of 2,9-disubstituted pyrrolopyrimidines 16–18.
Reagents and conditions: (a) 3,4,5-trimethoxyaniline (1.5 equiv), HCl
(1.0 equiv), n-BuOH, 110 °C, 4 h, 90%; (b) ArBr (1.5 equiv), 1,4-
dioxane, CuI (0.1 equiv), K3PO4 (2.0 equiv), trans-1,2-diaminocyclo-
hexane (0.1 equiv), 100 °C, 5 h, 40–85%.
Scheme 5. Synthesis of 22–25. Reagents and conditions: (a) (COCl)2,
DMF (cat), CH2Cl2, 25 °C, 3 h; (b) RNH2 (1.5 equiv), i-Pr2NEt
(20.0 equiv), 25 °C, 4 h, 45–85% for two steps.
Since a 2-pyridyl moiety at the 9-position is more
favorable over the corresponding phenyl substitution
at the same position,8 we next prepared a series of
analogs which contain both the 2-pyridinyl ring and
the 30-substitutions (28a–28h, Scheme 6). Coupling of
pyrrolopyrimidine intermediate 9 with 2-bromopyri-
dine 26, followed by hydrolysis of the resulting methyl
ester, furnished the corresponding carboxy analog 27
in good overall yield. The acid chloride was then
tributyltin azide (Scheme 4). Compared to the corre-
sponding nitriles, tetrazole 20 lost activity slightly, while
19 and 21 remained approximately equipotent.
A variety of amide analogs (22–25) were prepared from
the carboxy analogs 16i, 17g–17i by reaction with vari-
ous simple and functionalized amines (Scheme 5).
Among the ortho analogs (22a–22d), only the relatively
small methyl amide (22b) exhibited moderate activity.
While most of the analogs with substitutions at the 30
position (23–25) have reduced potency, a number of
them (23g–23i and 23k) retained their activity. Since
all of these more potent analogs contain basic pendant
amines of various chain lengths, it is likely that they
are directed towards solvent. While these modifications
did not improve the potency, they could serve as useful
handles to modulate the physicochemical properties of
these pyrrolopyrimidines at a later stage.
N
N
Br
N
H
N
HN
N
HN
N
a,b
+
N
N
CO2H
CO2Me
MeO
OMe
MeO
OMe
OMe
9
OMe
26
27
c,d
28a:
R=Me
R=iBu
R=(CH2)2NH2
R=(CH2)2NMe2
R=(CH2)3NH2
R=CH2CN
N
N
28b:
28c:
28d:
28e:
28f:
N
HN
N
N
H
N
N
N
N
CONHR
N
N
HN
N
HN
N
a
N
CN
( )n
28g:
28h:
R=(CH2)2CN
R=CH2CO2Me
MeO
OMe
3'( )n
3'
OMe
28
MeO
OMe
MeO
OMe
OMe
OMe
17c: n = 0
17d: n = 1
17e: n = 2
19: n = 0
20: n = 1
21: n = 2
Scheme 6. Synthesis of 28. Reagents and conditions: (a) 26 (1.2 equiv),
1,4-dioxane, CuI (0.1 equiv), K3PO4 (2.0 equiv), trans-1,2-diaminocy-
clohexane (0.1 equiv), 100 °C, 4 h, 60%; (b) NaOH (3.0 N), MeOH,
25 °C, 5 h, 85%; (c) (COCl)2, DMF (cat), CH2Cl2, 25 °C, 3 h; (d)
RNH2 (1.5 equiv), 25 °C, 2 h, 75–90%.
Scheme 4. Synthesis of 19–21. Reagents and conditions:
(a) n-Bu3SnN3, xylene, reflux, 20 h, 70–82%.

2692
H.-S. Choi et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2689–2692
formed from 27 using oxalyl chloride, and its reaction
with various amines produced the amide analogs 28a–
28h. Contrary to our expectations of simple additivity,
all these analogs had reduced FAK inhibitory activity.
One potential explanation is that an unfavorable con-
formation is induced through a lone-pair interaction
between the pyridyl nitrogen and adjacent carbonyl
oxygen.
inhibitory activity. Various acidic or basic moieties
could be incorporated onto the 9-aryl ring to mod-
ulate the interactions with the enzyme and potential-
ly their physicochemical properties. Our studies
suggest that the effect of the 2-pyridine nitrogen is
sensitive to its environment, and that the positioning
and orientation of a carboxy group are critical for
optimal interactions with the kinase. Future studies
will focus on optimizing the selectivity against other
kinases and the cellular activities of the pyrrolo[2,3-
d]pyrimidines.
In order to test this hypothesis, an analog containing
both a 2-pyridine ring and an extended carboxy group
(32) was synthesized from 7 (Scheme 7). To our delight,
compound 32 exhibits single-digit nanomolar IC50
against FAK and represents one of the most potent
FAK enzyme inhibitors reported to date. Molecular
modeling studies suggest that the carboxy group could
make a salt bridge with K454 (Fig. 3).
References and notes
1. Richardson, A.; Parsons, J. T. Nature 1996, 380, 538.
2. Gilmore, A. P.; Romer, L. H. Mol. Biol. Cell 1996, 7,
1209.
In summary, we have demonstrated how a structure-
based design approach can be used to design pyr-
rolo[2,3-d]pyrimidines which possess potent FAK
3. Hynes, R. O. Trends Cell Biol. 1999, 9, 33.
4. Ilic, D.; Furuta, Y.; Kanazawa, S.; Takeda, N.;
Sobue, K.; Nakatsuji, N.; Nomura, S.; Fukimoto, J.;
Okada, M.; Yamamoto, T.; Aizawa, S. Nature 1995,
377, 539.
5. Frisch, S. M.; Vuori, K.; Ruoslahti, E.; Chan-Hui, P. Y.
J. Cell Biol. 1996, 134, 793.
6. Miyazaki, T.; Kato, H.; Nakajima, M.; Sohda, M.; Fukai,
Y.; Masuda, N.; Manda, R.; Fukuchi, M.; Tsukada, K.;
Kuwano, H. Br. J. Cancer 2003, 89, 140.
Br
N
N
N
N
H
N
HN
N
HN
N
a
N
+
CO2tBu
MeO
OMe
MeO
OMe
CO2tBu
OMe
9
OMe
7. Mitra, S. K.; Hanson, D. A.; Schlaepfer, D. D. Nat. Rev.
Mol. Cell Biol. 2005, 6, 56.
29
30
b
8. Choi, H.-S.; Wang, Z.; Richmond, W.; He, X.; Yang, K.;
Jiang, T.; Sim, T.; Karanewsky, D.; Gu, X.-J.; Zhou, V.;
Liu, Y.; Ohmori, O.; Caldwell, J.; Gray, N.; He, Y. Bioorg.
Med. Chem. Lett. 2006, 16, 0000.
9. Lee, C. C., et al., manuscript in preparation.
10. Chari, R. V. J.; Kozarich, J. W. J. Org. Chem. 1982, 47,
2355.
N
N
N
N
HN
N
HN
N
c
N
N
CO2H
CO2H
MeO
OMe
MeO
OMe
OMe
OMe
32
31
11. Yu, M. S.; Lopez De Leon, L.; McGguire, M. A.; Botha,
G. Tetrahedron Lett. 1998, 39, 9347.
Scheme 7. Synthesis of 32. Reagents and conditions: (a) 29 (1.2 equiv),
1,4-dioxane, CuI (0.1 equiv), K3PO4 (2.0 equiv), trans-1,2-diaminocy-
clohexane (0.1 equiv), 100 °C, 4 h, 80%; (b) TFA, CH2Cl2, 25 °C, 3 h,
80%; (c) 10% Pd/C (cat), H2, EtOH, 25 °C, 8 h, 76%.
12. Seitz, D. E.; Lee, S.-H. Tetrahedron Lett. 1981, 22, 4909.
13. Ground state energies and conformations were calculated
using OPLS2003 forcefield with quantum mechanical ESP
charges. Quantum mechanical calculations were based on
DFT with 6-31G** basis set under continuum solvation
model (e = 80.4). The conformational energy penalty was
obtained from internal energy difference between the
bound and free forms of each ligand molecule including
solvation. Quantum mechanical charges were important
for proper assessments of energies since charge distribu-
tions had significant influence on desolvation energies.
Desolvation and torsional energy penalty were the two
most important factors that control the activity difference
among this class of compounds.
14. Computational docking: protein coordinates for docking
were obtained from the crystal structure of FAK in
complex with pyrrolopyrimidine 1. The binding site for
docking was defined as the collection of amino acids of
˚
FAK within 10 A radius centered on the bound ligand 1.
Waters and 1 were deleted from the pdb file and hydrogen
atoms were added to the amino acid residues. Compound
3D coordinates were generated using the MOE program.
Docking was performed using the default parameter
settings of GOLD v2.2 (Cambridge Crystallographic Data
Centre). All atom types and charges were assigned in
GOLD. 1,00,000 independent genetic algorithm (GA) runs
were performed for the ligand. The radius of the search
Figure 3. Docking of pyrrolopyrimidine 32 into the active site of
FAK.14
˚
was set to 10 A.