N-(3-(Phenylcarbamoyl)arylpyrimidine)-5-carboxamides as potent and selective inhibitors of Lck: Structure, synthesis and SAR

Article abstract of DOI:10.1016/j.bmcl.2007.11.123

N-3-(Phenylcarbamoyl)arylpyrimidine-5-carboxamides are a novel class of selective Lck inhibitors. This series of compounds derives its selectivity from a hydrogen bond with the gatekeeper Thr316 of the enzyme. X-ray co-crystal structural data, structure-activity relationships, and the synthesis of these inhibitors are reported herein.

Full text of DOI:10.1016/j.bmcl.2007.11.123

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 1172–1176
N-(3-(Phenylcarbamoyl)arylpyrimidine)-5-carboxamides as
potent and selective inhibitors of Lck: Structure, synthesis and SAR
Holly L. Deak,a,* John R. Newcomb,b Joseph J. Nunes,a Christina Boucher,b
Alan C. Cheng,c Erin F. DiMauro,a Linda F. Epstein,c Paul Gallant,b
Brian L. Hodous,a Xin Huang,c Josie H. Lee,b Vinod F. Patel,a
Stephen Schneider,b Susan M. Turcib and Xiaotian Zhuc
aDepartment of Medicinal Chemistry, Amgen Inc., One Kendall Square, Building 1000, Cambridge, MA 02139, USA
bDepartment of HTS & Molecular Pharmacology, Amgen Inc., One Kendall Square, Building 1000, Cambridge, MA 02139, USA
cDepartment of Molecular Structure, Amgen Inc., One Kendall Square, Building 1000, Cambridge, MA 02139, USA
Received 10 September 2007; revised 27 November 2007; accepted 30 November 2007
Available online 5 December 2007
Abstract—N-3-(Phenylcarbamoyl)arylpyrimidine-5-carboxamides are a novel class of selective Lck inhibitors. This series of com-
pounds derives its selectivity from a hydrogen bond with the gatekeeper Thr316 of the enzyme. X-ray co-crystal structural data,
structure–activity relationships, and the synthesis of these inhibitors are reported herein.
Ó 2007 Elsevier Ltd. All rights reserved.
T cell receptor (TCR) signal transduction pathways play a
role in the adaptive immune response through gene regu-
lation events, which lead to cytokine release, prolifera-
tion, and survival of antigen-specific T cells. A number
of protein kinases have been shown to be potentiators
of the TCR signal transduction pathways; among them
are Src family members, Lck (lymphocyte-specific kinase)
and Fyn.1 Mice and humans with Lck mutations exhibit
defective T cell maturation and signaling.2 This observa-
tion suggests that Lck inhibition may be a means of treat-
ing T cell-mediated autoimmune and inflammatory
diseases and transplant graft rejection.
While several groups3 have reported the discovery and
development of potent Lck inhibitors with efficacy in
in vivo models of inflammation, attaining selectivity
Figure 1. Representative example of 2-aminoquinazoline series (1).
over structurally related kinases has been an ongoing
challenge which has only recently been addressed.3b
cacy in a mouse model of inflammation. However, after
We recently reported a series of potent, nonselective
extensive SAR studies, it was possible to attain only
inhibitors of Lck, represented by 2-aminoquinazoline 1
(Fig. 1).3f The 2-aminoquinazolines exhibited excellent
potency, desirable pharmacokinetic properties, and effi-
through additional modifications to this scaffold.
modest selectivity over Tie-2, Jak3, KDR, and other ki-
nases. We believed further selectivity could be achieved
The X-ray co-crystal structure of 1 bound to Lck was
˚
Keywords: Lck inhibitor; Kinase inhibitor.
*
solved to 2.0 A resolution (Fig. 2). Binding of the inhib-
itor in the ATP-binding site induced the protein to as-
Corresponding author. Tel.: +1 617 444 5177; fax: +1 617 621
3907; e-mail: holly.deak@amgen.com
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.11.123

H. L. Deak et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1172–1176
1173
has threonine as its gatekeeper residue, which could
potentially limit selectivity. Herein, we describe the
structure–activity relationships and synthesis of N-(3-
(phenylcarbamoyl)aryl)pyrimidine-5-carboxamides,
novel class of selective inhibitors of Lck.
a
Compound 2 (Table 2) and subsequent compounds were
tested for Lck inhibition in a homogeneous time-re-
solved fluorescent (HTRF) kinase assay. For the pur-
pose of evaluating kinase selectivity, the compounds
were also tested against KDR, p38a, Jak3, and Tie-2.
Cellular activity was measured by testing the com-
pounds for the inhibition of T cell activation in a human
mixed lymphocyte reaction (MLR).5
While compound 2 exhibits modest potency in the Lck
enzyme inhibition assay (IC50 = 53 nM) and promising
selectivity (>50-fold) over KDR, p38a, Jak3, and Tie-2
(Table 2), it showed >10 lM cellular potency in our
MLR assay. We sought to improve the potency by syn-
thesizing variations of the trifluoromethylphenyl ring
that showed promise in our 2-aminoquinazoline SAR
studies.3f However, we found that none of these com-
pounds (exemplified by 3 and 4, Table 2) exhibited
sub-micromolar cellular activity.
Figure 2. Co-crystal structure of Lck with 1 (green). Oxygen atoms are
shown in red, and nitrogen atoms are blue.
sume the ‘DFG-out’ conformation.4 This orientation al-
lows the trifluoromethylphenyl ring to access an ex-
tended hydrophobic pocket. van der Waals
interactions between hydrophobic residues in this pock-
et and the aryl ring of the inhibitor are responsible for
the excellent potency observed in this series. Several
other interactions are visible in the co-crystal structure.
The 2-amino functional group and N3 of the quinazo-
line are involved in hydrogen bonds with Met319 in
the hinge region of the protein. The amide functionality
of 1 donates a hydrogen bond to Glu288 of the C-helix
and accepts an NH hydrogen bond from Asp382 of the
DFG sequence.
We further modified the scaffold to create additional
contacts with the enzyme in order to achieve greater en-
zyme and cellular potency. To engage the ribose pocket
of the enzyme, the methoxy group at the 4-position of
the pyrimidine was replaced by a phenoxy group (5, Ta-
ble 2). While compound 5 offered comparable enzyme
potency and maintained excellent selectivity, no increase
in cellular potency was realized.
Based on this structural information, we hypothesized
that selectivity could be gained by exploiting a sequence
difference between Lck and related kinases, particularly
KDR, Jak3, and Tie-2 (Table 1). We designed an inhib-
itor (2, Fig. 3) to be isosteric with the quinazolines but
containing an appropriately placed carbonyl to engage
in a hydrogen bond with the hydroxyl group of the
Thr316 gatekeeper residue in Lck. The gatekeeper resi-
dues of KDR, Jak3, and various other kinases are pre-
cluded from participating in this hydrogen bonding
interaction (Table 1). On the other hand, p38a also
We concurrently examined the effect of substitution on
the hinge-binding region of the pyrimidine. Functional-
ity at the 2-position of the pyrimidine should be easily
accommodated by the protein as evidenced by the co-
crystal structures of related compounds with Lck3f and
other kinases.6
Several amine-containing functional groups were incor-
porated into our scaffold at the 2-position of the pyrim-
idine. Varying alkyl chain lengths and amine structure
led to inhibitors 6–10 with Lck enzyme potency in the
range of 19–101 nM (Table 2). We also observed an in-
crease in cellular potency (MLR IC50 = 2.7 lM for 9)
through the addition of these groups. While attachment
of a 2-morpholinoethyl amine increased cellular potency
in 8 (MLR IC50 = 4.9 lM), no improvement was ob-
served in the corresponding 4-OPh analog 6 (MLR
IC50 = 10 lM). We were encouraged to find that incor-
poration of 4-(4-methylpiperazin-1-yl)aniline led to 11
which has greater potency in the Lck enzyme assay
(IC50 = 0.6 nM). This represents an 80-fold increase
over 2, while maintaining >400-fold selectivity over the
other enzymes in our kinase panel. Importantly, com-
pound 11 also exhibited vastly improved potency in
the MLR cellular assay (IC50 = 107 nM).
Table 1. Gatekeeper residues in Lck and other kinases
Lck
Thr
KDR
Val
p38a
Thr
Jak3
Met
Tie-2
Ile
We then designed compound 12, which presents an N–H
hydrogen bond donor to the gatekeeper, Thr316, in con-
Figure 3. Progression from quinazoline 1 to pyrimidine amide 2.

1174
H. L. Deak et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1172–1176
Table 2. SAR of pyrimidine amide seriesa
Compound R1
R2
R3
IC50 (nM)
Jak3
Lck KDR
p38a
Tie-2
MLR
2
3
4
5
NHCH3
NHCH3
NHCH3
NHCH3
H
CH3
CH3
CH3
Ph
53
7
8330
2720
>25,000 >25,000 >10,000
>25,000 >25,000 >10,000
>25,000 >10,000 >25,000 >25,000 >10,000
F
>25,000 254
CH3
H
44
25
>25,000 ND
>25,000 ND
>25,000 >25,000 >10,000
>25,000 >25,000 >10,000
6
7
8
H
H
H
Ph
41
CH3
CH3
101
22
>25,000 ND
>25,000 >25,000 ND
>25,000 >25,000 4920
2030
ND
9
CH3
CH3
CH3
CH3
19
45
>25,000 ND
>25,000 ND
>25,000 >25,000 2710
>25,000 >25,000 ND
10
11
12
F
CH3
0.6
84
255
856
2.4
334
3440
107
6679
>25,000 >25,000 ND
a IC50 values are means of two or more separate determinations, in duplicate.
trast to our other inhibitors (2–11) that present a hydro-
gen bond acceptor. Our objective was to determine
whether the enzyme would tolerate this alternative
hydrogen bonding partner. Compound 12 adopts its
conformation through rotation around the C5–C7
bond, and this new conformation is enforced by an
internal hydrogen bond between the carbonyl oxygen
and an aminomethyl group at the 4-position of the
pyrimidine. The effect of this modification on enzyme
potency is notable; while retaining potency on Lck, 12
proved to be a potent inhibitor of p38a (IC50 = 2 nM).
This observation can be explained by the structural dif-
ferences between the threonine gatekeeper residues of
p38a and Lck (vide infra).
Tie-2, and Jak3. The three-dimensional structures of
quinazoline 1 and pyrimidine 11 are similar, in which
the pyrimidine and pseudo-ring of 11 (generated
through intramolecular hydrogen bonding) occupy the
same space as the quinazoline of 1 (Fig. 4b). Another
important similarity between the structures of 1 and 11
in Lck is that the dihedral angle between the hinge-bind-
ing ring and the tolyl ring is 90°, which is the preferred
geometry based on the enzyme structure. Compounds 1
and 11 participate in the same hydrogen bonding inter-
actions with the hinge residue Met319 as well as Asp382
and Glu288. However, unlike quinazoline 1, pyrimidine
11 does not induce the kinase to adopt the ‘DFG out’
conformation. Instead, in this unusual binding mode,
the DFG motif may be prevented from adopting the
‘DFG out’ conformation because of interference with
the 4-methoxy group on the pyrimidine. In the ‘DFG
out’ conformation, the phenyl sidechain of Phe383
X-ray crystallographic studies contributed to our under-
standing of the selectivity observed in this new series
compared to the 2-aminoquinazoline series. A co-crystal
˚
˚
structure of Lck with 11 was solved to 2.3 A resolution
would be less than 3.0 A from the 4-methoxy group of
(Fig. 4a).7 As predicted, the pyrimidine amide oxygen is
involved in a hydrogen bond with the gatekeeper resi-
11. This interference may result in the displacement of
the DFG motif, thereby causing the sidechain of
Tyr360 to shift toward solvent. In this conformation,
Phe383 of the DFG motif is involved in a hydrophobic
˚
due, Thr316 (OH–O distance = 2.5 A), providing an
explanation for the selectivity observed over KDR,

H. L. Deak et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1172–1176
1175
imidine of 11 is involved in van der Waals interactions
with the residues succeeding the hinge region of Lck:
˚
with the a carbon of Gly322 (3.1 A) and with Tyr318
˚
˚
(3.2 A) and Met319 (3.0 A). Furthermore, the phenyl
ring of 11 also participates in a favorable hydrophobic
interaction with the Leu251 residue of the glycine-rich
loop. Due to similar interactions with the residues suc-
ceeding the hinge region, a degree of potency is regained
on tyrosine kinases such as Tie-2 and KDR. Conversely,
the analogous sequence in p38a (Met109-Gly110-
Ala111-Asp112) adopts a conformation that does not
interact favorably with the inhibitor due to the absence
of one amino acid (Gly). Therefore, compound 11
exhibits increased selectivity over p38a.
The synthesis of the 3-(phenylcarbamoyl)arylpyrimi-
dine-5-carboxamides is illustrated in Scheme 1.9 Com-
mercially available 4-methyl-3-nitrobenzoyl chloride 13
is treated with 2-fluoro-3-trifluoromethyl aniline in the
presence of base, followed by reduction of the nitro
group, to provide the substituted 3-amino-4-methyl-N-
phenylbenzamide (14). The aminobenzamide is then
reacted with the known 2,4-dichloropyrimidine-5-car-
bonyl chloride (15)10 in the presence of Hunig’s base
to furnish dichloropyrimidine 16. Addition of one
equivalent of sodium methoxide to the more reactive
4-position of the pyrimidine provides the penultimate
intermediate 17. Finally, 4-(4-methylpiperazin-1-yl)ani-
line is added to incorporate the amino hinge binding ele-
ment at the 2-position of the pyrimidine to yield
compound 11.
Figure 4. (a) Co-crystal structure of Lck with 11 (green). (b) Overlay of
1 and Lck (cyan) and 11 and Lck (magenta). Oxygen atoms are shown
in red and nitrogen atoms are blue.
edge-to-face interaction with the 2-fluoro-3-trifluoro-
˚
methyl phenyl ring of 11 (3.4 A).
The observation that compounds 2–11 are selective over
p38a, despite its threonine gatekeeper residue, can be ex-
plained by examining co-crystal structures of inhibitors
bound to p38a. In Lck, the Thr316 sidechain is posi-
tioned to donate a hydrogen bond to the inhibitor
through the c-OH. However, published structures of
p38a show that the gatekeeper threonine hydroxyl
group is involved in a bridging hydrogen bond network
with main chain atoms from strands b5 and b6.8 In this
conformation, the gatekeeper threonine is available to
engage a hydrogen bond donor, but not a hydrogen
bond acceptor, such as compound 11. This argument
also applies to the potency on p38a observed for com-
pound 12, since the threonine oxygen lone pairs can
form a hydrogen bond with the NH of the pyrimidine
amide.
Scheme 1. Preparation of 3-(phenylcarbamoyl)arylpyrimidine-5-car-
boxamide 11. Reagents and conditions: (a) 2-fluoro-3-trifluoromethyl
aniline (1.05 equiv), DIPEA (2.0 equiv), THF (0.30 M), rt; (b) H2
(1 atm), Pd/C (30% by wt), MeOH (0.10 M), rt, 64% yield (over two
steps); (c) 15 (1.1 equiv), DIPEA (2.0 equiv), EtOAc (0.50 M), rt, 82%
yield; (d) NaOMe (0.95 equiv), CH3CN (0.10 M), 0 °C; (e) 4-(4-
methylpiperazin-1-yl)aniline (1.1 equiv), TFA (2.0 equiv), THF/DMF
(5:1, 0.17 M), 70 °C, 13% yield (over two steps).
X-ray crystallographic studies also assisted us in
accounting for the increase in potency we observe for
compound 11 compared to other compounds in this ser-
ies (Fig. 4a). The phenyl ring appended to the aminopyr-

1176
H. L. Deak et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1172–1176
Gallant, P.; Gore, A.; Gu, Y.; Hsieh, F.; Huang, X.; Lee,
J. H.; Metz, D.; Middleton, S.; Mohn, D.; Morgenstern,
K.; Morrison, M. J.; Novak, P. M.; Oliveira-dos-Santos,
A.; Powers, D.; Rose, P.; Schneider, S.; Sell, S.; Tudor, Y.;
Turci, S. M.; Welcher, A.; White, R. D.; Zack, D.; Zhao,
H.; Zhu, L.; Zhu, X.; Ghiron, C.; Amouzegh, P.; Ermann,
M.; Jenkins, J.; Johnston, D.; Napier, S.; Power, E.
J. Med. Chem. 2006, 49, 4981; (f) 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.; Epstein, L. F.; Faust, T.; Gallant, P.; Geuns-
Meyer, S. D.; Gore, A.; Gu, Y.; Henkle, B.; Hodous, B.
L.; Hsieh, F.; Huang, X.; Kim, J. L.; Lee, J. H.; Martin,
M. W.; Masse, C. E.; McGowan, D. C.; Metz, D.; Mohn,
D.; Morgenstern, K. A.; Oliveira-dos-Santos, A.; Patel, V.
F.; Powers, D.; Rose, P. E.; Schneider, S.; Tomlinson, S.
A.; Tudor, Y.-Y.; Turci, S. M.; Welcher, A. A.; White, R.
D.; Zhao, H.; Zhu, L.; Zhu, X. J. Med. Chem. 2006, 49,
5671.
In conclusion, we have discovered a new series of potent
inhibitors of Lck with enhanced selectivity in compari-
son to their quinazoline predecessors, based on a ratio-
nally designed hydrogen bonding interaction with the
gatekeeper threonine residue of the enzyme. We also
found that aryl functionality at the 2-position of the
pyrimidine can provide much improved enzyme and
cellular potency. This study has provided us with a
promising lead molecule 11 on which further studies
can be based. In addition, the discovery that 11 induces
a unique conformation of the DFG sequence of Lck
may lead to new opportunities for inhibitor design.
References and notes
1. (a) Marth, J. D.; Lewis, D. B.; Cooke, M. P.; Mellins, E.
D.; Gearn, M. E.; Samelson, L. E.; Wilson, C. B.; Miller,
A. D.; Perlmutter, R. M. J. Immunol. 1989, 142, 2430; (b)
Groves, T.; Smiley, P.; Cooke, M. P.; Forbush, K.;
Perlmutter, R. M.; Guidos, C. J. Immunity 1996, 5, 417; (c)
Palacios, E. H.; Weiss, A. Oncogene 2004, 23, 7990.
4. Liu, Y.; Gray, N. S. Nat. Chem. Biol. 2006, 2, 358, and
references therein.
5. For descriptions of assays, including the human mixed
lymphocyte reaction, please see Ref. 3e.
2. (a) Molina, T. J.; Kishihara, K.; Siderovskid, D. P.; van
Ewijk, W.; Narendran, A.; Timms, E.; Wakeham, A.;
Paige, C. J.; Hartmann, K.-U.; Veillette, A.; Davidson, D.;
Mak, T. W. Nature 1992, 357, 161; (b) Straus, D. B.;
Weiss, A. Cell 1992, 70, 585; (c) Levin, S. D.; Anderson, S.
J.; Forbush, K. A.; Perlmutter, R. M. EMBO J. 1993, 12,
1671; (d) Goldman, F. D.; Ballas, Z. K.; Schutte, B. C.;
Kemp, J.; Hollenback, C.; Noraz, N.; Taylor, N. J. Clin.
Invest. 1998, 102, 421; (e) Hubert, P.; Bergeron, F.;
Ferreira, V.; Seligmann, M.; Oksenhendler, E.; Debre, P.;
Autran, B. Int. Immunol. 2000, 12, 449; (f) Seddon, B.;
Zamoyska, R. J. Immunol. 2002, 169, 2997.
3. (a) Snow, R. J.; Cardozo, M. G.; Morwick, T. M.;
Busacca, C. A.; Dong, Y.; Eckner, R. J.; Jacober, S.;
Jakes, S.; Kapadia, S.; Lukas, S.; Panzenbeck, M.; Peet,
G. W.; Peterson, J. D.; Prokopowicz, A. S., III; Sellati, R.;
Tolbert, R. M.; Tschantz, M. A.; Moss, N. J. Med. Chem.
2002, 45, 3394; (b) Das, J.; Lin, J.; Moquin, R. V.; Shen,
Z.; Spergel, S. H.; Wityak, J.; Doweyko, A. M.; DeFex, H.
F.; Fang, Q.; Pang, S.; Pitt, S.; Shen, D. R.; Schieven, G.
L.; Barrish, J. C. Bioorg. Med. Chem. Lett. 2003, 13, 2145;
(c) Burchat, A.; Borhani, D. W.; Calderwood, D. J.; Hirst,
G. C.; Li, B.; Stachlewitz, R. F. Bioorg. Med. Chem. Lett.
2006, 16, 118; (d) Chen, P.; Doweyko, A. M.; Norris, D.;
Gu, H. H.; Spergel, S. H.; Das, J.; Moquin, R. V.; Lin, J.;
Wityak, J.; Iwanowicz, E. J.; McIntyre, K. W.; Shuster, D.
J.; Behnia, K.; Chong, S.; de Fex, H.; Pang, S.; Pitt, S.;
Shen, D. R.; Thrall, S.; Stanley, P.; Kocy, O. R.; Witmer,
M. R.; Kanner, S. B.; Schieven, G. L.; Barrish, J. C.
J. Med. Chem. 2004, 47, 4517; (e) Martin, M. W.;
Newcomb, J.; Nunes, J. J.; McGowan, D. C.; Armistead,
D. M.; Boucher, C.; Buchanan, J. L.; Buckner, W.; Chai,
L.; Elbaum, D.; Epstein, L. F.; Faust, T.; Flynn, S.;
6. (a) Fraley, M. A.; Hoffman, W. F.; Arrington, K. L.;
Hungate, R. W.; Hartman, G. D.; McFall, R. C.; Coll, K.
E.; Rickert, K.; Thomas, K. A.; McGaughey, G. B. Curr.
Med. Chem. 2004, 11, 709; (b) Heron, N. M.; Anderson,
M.; Blowers, D. P.; Breed, J.; Eden, J. M.; Green, S.; Hill,
G. B.; Johnson, T.; Jung, F. H.; McMiken, H. H.;
Mortlock, A. A.; Pannifer, A. D.; Pauptit, R. A.; Pink, J.;
Roberts, N. J.; Rowsell, S. Bioorg. Med. Chem. Lett. 2006,
16, 1320.
7. The kinase domain of human Lck (residues 225–501) with
a C-terminal His6 tag was expressed in insect cells and
purified by immobilized metal affinity, anion exchange,
and size exclusion chromatographies. The protein was
then phosphorylated by incubation with 5 mM Mg++ÆATP
for 10 min at room temperature. Phosphorylated Lck was
further purified by anion exchange chromatography and
concentrated to 10 mg/mL. The coordinates for the X-ray
co-crystal structure of Lck and 11 have been deposited in
the PDB. The RCSB ID code is RCSB045010 and the
PDB ID code is 3B2W.
8. For examples, see: (a) Tamayo, N.; Liao, L.; Goldberg,
M.; Powers, D.; Tudor, Y.-Y.; Yu, V.; Wong, L. M.;
Henkle, B.; Middleton, S.; Syed, R.; Harvey, T.; Jang, G.;
Hungate, R.; Dominguez, C. Bioorg. Med. Chem. Lett.
2005, 15, 2409; (b) Pargellis, C.; Tong, L.; Churchill, L.;
Cirillo, P. F.; Gilmore, T.; Graham, A. G.; Grob, P. M.;
Hickey, E. R.; Moss, N.; Pav, S.; Regan, J. Nat. Struct.
Biol. 2002, 9, 268.
1
9. All compounds were characterized by H NMR and LC/
MS and their purity determined to be >95% by reverse
phase HPLC.
10. Luecking, U.; Siemeister, G.; Schaefer, M.; Briem, H. Ger.
Offen. DE 10239042, 2004.