Trifluoromethylpyrimidine-based inhibitors of proline-rich tyrosine kinase 2 (PYK2): Structure-activity relationships and strategies for the elimination of reactive metabolite formation

Article abstract of DOI:10.1016/j.bmcl.2008.10.030

The synthesis and SAR for a series of diaminopyrimidines as PYK2 inhibitors are described. Using a combination of library and traditional medicinal chemistry techniques, a FAK-selective chemical series was transformed into compounds possessing good PYK2 potency and 10- to 20-fold selectivity against FAK. Subsequent studies found that the majority of the compounds were positive in a reactive metabolite assay, an indicator for potential toxicological liabilities. Based on the proposed mechanism for bioactivation, as well as a combination of structure-based drug design and traditional medicinal chemistry techniques, a follow-up series of PYK2 inhibitors was identified that maintained PYK2 potency, FAK selectivity and HLM stability, yet were negative in the RM assay.

Full text of DOI:10.1016/j.bmcl.2008.10.030

Bioorganic & Medicinal Chemistry Letters 18 (2008) 6071–6077
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Trifluoromethylpyrimidine-based inhibitors of proline-rich tyrosine kinase
2 (PYK2): Structure–activity relationships and strategies for the elimination
of reactive metabolite formation
*
Daniel P. Walker , F. Christopher Bi, Amit S. Kalgutkar, Jonathan N. Bauman, Sabrina X. Zhao, John R. Soglia,
Gary E. Aspnes, Daniel W. Kung, Jacquelyn Klug-McLeod, Michael P. Zawistoski, Molly A. McGlynn,
Robert Oliver, Matthew Dunn, Jian-Cheng Li, Daniel T. Richter, Beth A. Cooper, John C. Kath,
Catherine A. Hulford, Christopher L. Autry, Michael J. Luzzio, Ethan J. Ung, W. Gregory Roberts,
Peter C. Bonnette, Leonard Buckbinder, Anil Mistry, Matthew C. Griffor, Seungil Han, Angel Guzman-Perez
Pfizer Global Research and Development, Eastern Point Road, Groton, CT 06340, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
The synthesis and SAR for a series of diaminopyrimidines as PYK2 inhibitors are described. Using a com-
bination of library and traditional medicinal chemistry techniques, a FAK-selective chemical series was
transformed into compounds possessing good PYK2 potency and 10- to 20-fold selectivity against FAK.
Subsequent studies found that the majority of the compounds were positive in a reactive metabolite
assay, an indicator for potential toxicological liabilities. Based on the proposed mechanism for bioactiva-
tion, as well as a combination of structure-based drug design and traditional medicinal chemistry tech-
niques, a follow-up series of PYK2 inhibitors was identified that maintained PYK2 potency, FAK selectivity
and HLM stability, yet were negative in the RM assay.
Received 7 September 2008
Revised 7 October 2008
Accepted 7 October 2008
Available online 11 October 2008
Keywords:
PYK2
FAK
Ó 2008 Elsevier Ltd. All rights reserved.
Kinase
Osteoporosis
Bioactivation
Reactive metabolite
Proline-rich tyrosine kinase 2 (PYK2) and focal adhesion kinase
(FAK) are nonreceptor tyrosine kinases, and together they constitute
the FAK subfamily.1 Unlike FAK expression, which is ubiquitous,
PYK2 expression is relatively restricted, with highest levels in the
brain and the hematopoietic system. In contrast to the embryonic
lethality in FAK-deficient mice, PYK2 À/À mice are viable and fer-
tile.2 In a prior study, we found that the skeletal phenotype of adult
PYK2 À/À female mice has significantly increased bone mass and
bone mineral density compared to wild-type mice.3 Furthermore,
daily oral administration of a small-molecule PYK2 inhibitor (PF-
431396, 1a) resulted in an increase in bone formation and protection
against bone loss in the ovariectomized (OVX) rat model, an estab-
lished preclinical model of postmenopausal osteoporosis.3 The fact
that this efficacious response resulted from increased bone forma-
tion with minimal change in bone resorption has important thera-
peutic implications. Oral agents approved to treat osteoporosis
include bisphosphonates, estrogen, and selective estrogen receptor
modulators (SERMs), all of which are antiresorptive in nature and
are insufficient for restoring bone in critically osteoporotic patients.
Parathyroid hormone 1–34 is the only approved bone anabolic
agent; however, it is an injectable that increases bone turnover with
a balance favoring bone formation. Considering the usual challenges
associated with protein-based therapeutics, such as inconvenient
routes of administration, an orally bioavailable small-molecule
PYK2 inhibitor represents an attractive anabolic therapy for patients
afflicted by low bone mass and offers a new opportunity to impact
this unmet medical need in the aging population.
H
CF3
NH
N
N
O
N
N
H
Me
PF-431396 (1a)
N
S
O
O
Me
Pyrimidine 1a was an attractive compound due to its in vitro PYK2
potency (PYK2 IC50 = 31 nM) and in vivo efficacy.3 However, subse-
quent studies accessing the bioactivation potential of 1a and struc-
* Corresponding author. Tel.: +1 636 247 1192.
E-mail address: daniel.p.walker@pfizer.com (D.P. Walker).
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.10.030

6072
D. P. Walker et al. / Bioorg. Med. Chem. Lett. 18 (2008) 6071–6077
Not surprisingly, there was as well significant overlap in the
Method A (X = CF3)
X
5
X
-or-
N
N
SAR for PYK2 and FAK at the C4-position of the pyrimidine ring.
This is likely due to the high degree of homology between PYK2
and FAK within the ATP binding pocket.8 Nevertheless, some differ-
ences in the SARs were identified. For instance, whereas unsubsti-
tuted benzylamine 1j showed a modest preference for FAK, meta-
substitution significantly increased FAK activity and selectivity
(compounds 1k and 1l). Conversely, ortho-substitution appeared
to result in a modest selectivity for PYK2 (compounds 1m and
1n). para-Substitution resulted in decreased activity for both FAK
and PYK2 (data not shown). In an effort to identify compounds
with better HLM stability and PYK2 selectivity, we prepared a
diversity library possessing aliphatic C4 amines. Two compounds
emerged from these efforts, caprolactam 1o and cyclopentanedi-
amine 1p, that showed good PYK2 potency and 6- to 10-fold selec-
tivity against FAK. Cyclopentanediamine 1p was further attractive
in that it appeared to be stable in HLM towards oxidative
metabolism.
During the course of the SAR studies to optimize PYK2 inhibi-
tory potency and selectivity, it was observed that our early lead
compound 1a was positive in a high-throughput screen (HTS) for
reactive metabolite formation. This screen examines the bioactiva-
tion potential of drug candidates in HLM via the detection of GSH
and/or GSH-EE-captured reactive metabolites.9 There are examples
of drugs that are false positives in the HTS RM assay, yet are not
associated with a significant incidence of toxicity.10 However, since
our interest was to develop a drug candidate for treating osteopo-
rosis, which is a nonlife threatening disease, we felt it was impor-
tant to eliminate the bioactivation liability within this chemical
series. Liquid chromatography–tandem mass spectrometry (LC/
MS/MS) analysis of the NADPH-supplemented HLM incubations
containing 1a and GSH or GSH-EE9,11,12 led to the detection of sul-
fydryl conjugates with molecular ions (MH+) at 812 (GSH) and 840
(GSH-EE), respectively. The formation of these conjugates was
NADPH-dependent suggesting the involvement of cytochrome
P450 in the bioactivation of 1a. Also, inclusion of the specific
P4503A4 inhibitor ketoconazole in the HLM incubations elimi-
nated conjugate formation and overall metabolism implicating
that P4503A4 was responsible for the oxidative metabolism/bioac-
tivation of 1a. Furthermore, in both cases, additional chromato-
graphic separation revealed the presence of two isomeric peaks
as depicted with the GSH conjugate of 1a (Fig. 1, panel A). Overall,
the molecular mass of the conjugate was consistent with the addi-
tion of one molecule of GSH or GSH-EE to the molecular mass of
the parent compound 1a.
4
2
Ar
R
N
N
N
H
Cl
N
Cl
Method B
(X = H, Me, CN, Cl, Br)
H
1a-r
2a-n
Scheme 1. General synthetic route for the preparation of pyrimidines of types 1
and 2. Method A reagents and conditions: (a) 2.0 equiv, ZnCl2 (1.0 M solution in
ether), CH2Cl2–tBuOH (1:1), À5 °C, 1 h; ArNH2, 1.1 equiv Et3N, À5 ? 40 °C, 45–95%;
(b) RNH2, DMF, Na2CO3, 85 °C, 1–2 h, 60–95%. Method B reagents and conditions: (a)
RNH2, i-Pr2NEt, THF, room temperature, 24 h, 75–95%; (b) ArNH2, i-Pr2NEt, dioxane,
110 °C, 24–48 h, 10–50%.
turally-related analogs in NADPH- and glutathione-ethyl ester
(GSH-EE)-supplemented human liver microsomes (HLM) revealed
the formation of sulfydryl conjugates, a phenomenon consistent
with reactive metabolite formation. With an increasing awareness
of potential toxicological liabilities that can occur via bioactivation
of toxicophores within drug candidates,4 we felt that it was impor-
tant to identify the metabolic pathway responsible for reactive
metabolite formation and to eliminate this undesirable attribute
via rational structural modifications. We detail below the synthesis
and structure–activity relationships (SARs) for a series of diamino
pyrimidines as PYK2 inhibitors. We also detail our efforts to charac-
terize the structure of the GSH conjugate of 1a. This information
provided insights into the structure of the reactive metabolite and
the bioactivation pathway responsible for its formation thus allow-
ing chemical intervention to eliminate the potential safety hurdle.
Pyrimidines of types 1 and 2 were prepared in two steps accord-
ing to Scheme 1. Method A was utilized for pyrimidines possessing
a trifluoromethyl group at C5,5,6 whereas all other pyrimidines
examined in this study were prepared by Method B.
Early in-house SAR studies on the pyrimidine series against the
related kinase, FAK, identified 5-aminooxindole as a suitable C2
group (Table 1).7 While other 6,5-fused heterocycles, such as in-
dole 1b and indazole 1c, were also tolerated at this position, these
heterocycles exhibited sub-optimal oral pharmacokinetics in the
rat (oral bioavailability < 5%) due to a combination of poor aqueous
solubility and first pass oxidative metabolism. SAR studies also
identified trifluoromethyl as a suitable C5 group (compounds 1d–
1i). While the role of the C5 substituent in the inhibitory process
is unclear, it is possible that the C5 substituent modulates the acid-
ity of the C2 N–H bond, which is important for ligand binding to
the kinase (vide infra). Further optimization of the pyrimidine ser-
ies for PYK2 activity also found that 5-aminooxindole and trifluo-
romethyl were optimal C2 and C5 groups, respectively.
H
CF3
CF3
NH
N
N
N
N
O
P4503A4
O
N
N
NH
N
N
H
H3C
O
1a
N
S
CH3
O
H
N
CF3
NH
N
O
O
N
H
N
GSH
S
O
H3C
O
N
N
S
H
H2N
HN
CH3
O
COOH
COOH
Scheme 2. Proposed mechanism of P450-catalyzed bioactivation of 1a.

D. P. Walker et al. / Bioorg. Med. Chem. Lett. 18 (2008) 6071–6077
6073
Table 1
In vitro IC50 values for FAK and PYK2 inhibition, HLM stability and propensity for reactive metabolite formation by compounds 1a–1r
Y
N
Z
X
N
N
N
H
H
1a-r
HLM Ehc
0.78
RMAd,e
a
b
Compound
X
Y
Z
FAK IC50 (nM)
PYK2 IC50 (nM)
Me
SO2Me
H
N
N
1a PF-431396
CF3
1.5 (n = 3)
11 (n = 3)f
Positive
O
N
N
N
1b
1c
CF3
CF3
160
16
1200
600
>0.85
0.55
Positive
N.T.
H
N
N
H
N
N
N
N
N
N
N
1d
1e
1f
H
>10,000
2200
>10,000
>3000
>1000
580
<0.26
<0.30
<0.28
0.44
Positive
Positive
Positive
N.T.
O
O
O
O
O
O
O
H
N
Me
Cl
H
N
77 (n = 3)
46 (n = 5)
190
H
N
1g
1h
1i
Br
H
N
CN
CF3
CF3
1800
680
<0.26
<0.26
<0.26
Positive
Positive
N.T.
H
N
19
H
N
1j
34
120
H
N
1k
1l
CF3
CF3
CF3
CF3
CF3
CF3
1.9 (n = 7)
9.0 (n = 9)
59
190 (n = 7)
>1100
15
0.59
0.35
0.55
>0.93
0.55
<0.26
Positive
Positive
Positive
Positive
Positive
Positive
O
O
O
O
O
O
N
SO2Me
H
N
SO2Me
SO2Me
H
N
1m
1n
1o
1p
H
N
O2S N
170 (n = 4)
17 (n = 4)
H
H
N
N
O
300
49
H
N
810, 1600
53
NMe2
(continued on next page)

6074
D. P. Walker et al. / Bioorg. Med. Chem. Lett. 18 (2008) 6071–6077
Table 1 (continued)
HLM Ehc
0.29
RMAd,e
a
b
Compound
X
Y
Z
FAK IC50 (nM)
PYK2 IC50 (nM)
210
H
N
N
O
1q
CF3
38
Positive
H
N
O
1r
CF3
N.T.
33
N.T.
Positive
a
FAK ELISA protocol. IC50 values represent the concentration required to inhibit 50% of poly-Glu-Tyr phosphorylation by GST:FAK (see: Ref. 19). Numbers indicate IC50
values generated from individual 6-point concentration response relationships in triplicate.
b
PYK2 ELISA protocol. IC50 values represent the concentration required to inhibit 50% of poly-Glu-Tyr phosphorylation by GST:PYK2 (see: Ref. 19). Numbers indicate IC50
values generated from individual 6-point concentration response relationships in triplicate.
c
Eh = in vitro hepatic extraction ratio, which is obtained from dividing estimated hepatic blood clearance of test compounds by the human hepatic blood flow of 20 mL/
min/kg. Protocols for measuring half-lives in HLM and subsequent scaling to blood clearance have been published (see: Ref. 20).
d
For compound 1b, the molecular mass of the GSH-EE conjugate corresponded to addition of the thiol to a N-demethylated metabolite of the parent compound.
For all other analogs, the molecular mass of the GSH-EE conjugate corresponded to addition of the thiol to the parent compound suggesting a similar bioactivation
e
mechanism as shown for 1a.
f
Buckbinder et al. (see: Ref. 3) reports an IC50 of 31 nM for compound 1a using a fluorescence polarization assay.
The product ion spectrum obtained by collision-induced disso-
ciation (CID) of the MH+ at m/z 812 for the isomeric GSH conjugates
of 1a produced identical fragment ions at m/z 724, 683, 623, 551,
539, and 198 (Fig. 1, panel B). The presence of the fragment ion
at m/z 19813 established that the C4 substituent was unaltered
(see Fig. 1, panel B), indicating that GSH conjugation had occurred
on the 5-aminooxindole or the trifluoromethylpyrimidine ring. The
fragment ion at m/z 683 corresponded with the loss of the pyroglu-
tamate moiety, which represents a characteristic fragment ion de-
rived from GSH conjugates.14 The fragment ion at m/z 539 was
assigned as a cleavage adjacent to the cysteinyl thioether moiety
with charge retention on the 1a residue. The occurrence of the
fragment ion at m/z 539 is consistent with the presence of an aro-
matic thioether motif in 1a.14 A proposed structure for the GSH
conjugate(s) of 1a that is consistent with the observed mass spec-
trum, is shown in Figure 1, panel B. The assigned regiochemistry of
GSH addition as shown is arbitrary and for illustrative purposes
only. A mechanism for its formation is shown in Scheme 2. Thus,
P450-catalyzed two-electron oxidation of the 5-aminooxindole
motif would generate the electrophilic bis-imine followed by Mi-
chael addition of GSH. A similar mechanism has recently been re-
ported for a series of aryl amide tyrosine kinase inhibitors.15 In
general, this bioactivation sequence is analogous to the one ob-
served with the prototypic hepatotoxin and anti-inflammatory
agent acetaminophen.16 Not surprisingly, all analogs possessing
the 5-aminooxindole group at C2 were positive in the reactive
metabolite assessments (see Table 1). Likewise, analogs 1q and
1r, in which the oxindole group was replaced with the correspond-
ing dihydroquinolinone and tetrahydrobenzazepinone functional-
ities, respectively, also formed sulfydryl conjugates (Table 1). The
molecular masses of these conjugates suggested that a similar bio-
activation pathway involving the para-aminophenylacetamido
framework was operative.
moiety, in 1a with an electron withdrawing group should lower
the compound’s susceptibility to undergo CYP-mediated bioactiva-
tion. For this exercise, we limited the C5 substituent to trifluoro-
methyl and the C4 substituent to either cyclopentanediamine or
N-methyl-methanesulfonamidobenzylamine since it was unlikely
that these groups contribute significantly to the bioactivation
process.
Compounds possessing modifications to the C2 5-aminooxin-
dole moiety are shown in Table 2. We were pleased to discover that
analogs possessing a variety of aryl amines at C2 were negative in
the RM assay. Furthermore, several of the compounds showed
PYK2 activity and FAK selectivity similar to their corresponding
5-aminooxindole counterparts. Thus, isoindolinone 2a, where the
carbonyl and amino groups of the lactam ring of 1a have been re-
versed, displayed similar PYK2 potency and FAK selectivity to that
of 1a. Tetrahydrobenzazepinone 2b was negative in the RM assay,
but it was 3-fold less potent against PYK2 and was less stable in
HLM. Both dihydrobenzothiophene-S,S-dioxide 2c and dihydro-
benzisothiazole-S,S-dioxide 2d showed similar PYK2 activity and
HLM stability compared to the corresponding oxindole analog 1p.
In general, the 6,5-fused heterocycles, where an electron with-
drawing group (EWG) is directly attached to the ring, were effec-
tive replacement groups for the oxindole. Analogs 2e–2i, which
possess a directly linked EWG but do not contain a fused ring, also
eliminated the bioactivation potential. The PYK2 potency and HLM
stability of these analogs was in-line with the corresponding fused
ring compounds. Sulfone 2e displayed good stability in HLM,
whereas sulfonamide 2g and carboxamides 2h and 2i were less
stable than their fused ring counterparts.
Azanorbornene 2j is an example of an analog that does not pos-
sess a directly linked EWG yet was negative in the RM assay. Inter-
estingly, this analog was 9-fold more potent against PYK2 than the
corresponding oxindole analog 1p. It is possible the acetamide car-
bonyl on the azanorbornene is making a favorable interaction out
of the plane of the benzene ring with PYK2, which would not be
possible with the planar oxindole. Pyridine 2k and imidazopyri-
dine 2l represent two additional examples that were negative in
the RM assay yet do not possess a directly linked EWG. In these
cases, it is likely the pyridine nitrogen has a significant effect on
the electron density of the aryl ring such that a bioactivation mech-
anism similar to 1a is not favored.
Our strategy was to eliminate the bioactivation liability ob-
served with 1a while maintaining or improving PYK2 potency,
FAK selectivity and HLM stability. The X-ray co-crystal structure
of pyrimidine 1n bound to the ATP binding pocket of PYK2 is
detailed in Figure 2.17 The structure shows N1 of the pyrimidine
and the C2 amino group making key hydrogen bonds to the hinge
residue (Tyr505). More importantly, the structure shows the
5-membered ring lactam of the oxindole partially solvent exposed.
Thus, it was thought that making small structural modifications to
the lactam ring should be possible without adversely affecting
PYK2 activity. Furthermore, based on the propensity of electron-
rich aryl rings to undergo bioactivation by CYP enzymes,18 we
reasoned that replacing the fused lactam, an electron donating
While it was reasonable to anticipate lack of GSH conjugates for
compounds 2a–2l due to disruption of the electron-rich bis-aniline
architecture, we were surprised to find that benzosultamderivatives
2m and 2n were devoid of sulfydryl conjugate formation in the HTS
reactive metabolite screen as well as in separate HLM incubations

D. P. Walker et al. / Bioorg. Med. Chem. Lett. 18 (2008) 6071–6077
6075
Table 2
In vitro IC50 values for FAK and PYK2 inhibition, HLM stability and propensity for reactive metabolite formation by compounds 2a–2n
CF3
Z1-2
SO2Me
N
N
Z1 =
Z2 =
X
N
N
N
N
H
H
2a-n
a
b
Compound
X
Z
FAK IC50 (nM)
PYK2 IC50 (nM)
In vitro HLM Ehc
N.T.
RMAd
O
2a
1
2
0.50
5.0
Negative
HN
HN
O
2b
880
143
0.59
Negative
O
S
O
2c
2d
2e
2f
2
2
2
1
623
67 (n = 3)
25, 65
0.37
0.44
<0.28
0.93
Negative
Negative
Negative
Negative
O
S
O
3200
N
O
O
O
S
1300, 2000
0.94
95 (n = 4)
35, 66
O
S
S
O
O
2g
2h
2
2
3800, 4500e
154f
0.66
0.66
Negative
Negative
N
H
O
3300 (n = 4)e
72 (n = 4)f
Et2N
Me2N
2i
2
2
2
3900, 4000e
99hf
0.66
0.63
0.48
Negative
Negative
Negative
O
N
Ac
2j
72
6
N
2k
4100, 6900e
430f
N
2l
2
1
N.T.
150f
0.41
0.71
Negative
Negative
N
H
N
O
2m
1.0, 5.0
2.0, 3.0
S
O
(continued on next page)

6076
D. P. Walker et al. / Bioorg. Med. Chem. Lett. 18 (2008) 6071–6077
Table 2 (continued)
In vitro HLM Ehc
<0.26
RMAd
a
b
Compound
X
Z
2
FAK IC50 (nM)
PYK2 IC50 (nM)
H
N
O
2n
>650 (n = 5)
50 (n = 5)
Negative
S
O
a
FAK ELISA protocol. IC50 values represent the concentration required to inhibit 50% of poly-Glu-Tyr phosphorylation by GST:FAK (see: Ref. 19). Numbers indicate IC50
values generated from individual 6-point concentration response relationships in triplicate.
b
PYK2 ELISA protocol. IC50 values represent the concentration required to inhibit 50% of poly-Glu-Tyr phosphorylation by GST:PYK2 (see: Ref. 19). Numbers indicate IC50
values generated from individual 6-point concentration response relationships in triplicate.
c
Eh = hepatic extraction ratio, which is obtained from dividing estimated hepatic blood clearance of test compounds by the human hepatic blood flow of 20 mL/min/kg.
Protocols for measuring half-lives in HLM and subsequent scaling to blood clearance have been published (see: Ref. 20).
d
Reactive metabolite assessments in the HTS screening mode were conducted using GSH-EE as trapping agent in NADPH-supplemented HLM.
FAK fluorescence polarization (FP) protocol as described in Ref. 3 using FAK kinase domain construct as described in Ref. 19 IC50 values represent the concentration to
e
inhibit 50% of the phosphorylation of a peptide substrate relative to vehicle control. Numbers indicate IC50 values generated from individual 8-point concentration response
relationships in triplicate.
f
PYK2 fluorescence polarization (FP) protocol as described in Ref. 3 IC50 values represent the concentration to inhibit 50% of the phosphorylation of a peptide substrate
relative to vehicle control. Numbers indicate IC50 values generated from individual 8-point concentration response relationships in triplicate.
with GSH. The precise reason(s) for lack of GSH conjugate formation,
despite the presence of the para-aminophenylsulfonamido group,
remains unclear and is the subject of an independent investigation
in our laboratory. In addition to the lack of reactive metabolite liabil-
ity, both benzosultams 2m and 2n show good PYK2 potency. Benzo-
sultam 2n was further attractive in that it showed 20-fold FAK
selectivity and was stable in HLM.
In summary, we have described the synthesis and SAR for a
series of diaminopyrimidines as PYK2 inhibitors. Using a combi-
nation of library and traditional medicinal chemistry techniques,
14.45
3200
2800
14.04
2400
2000
1600
1200
800
400
15.66
0
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time (min)
O
683.2
O
H
N
CF3H3C
S
100
N
N
CH3
O
N
N
N
H
H
S
539 (+ 2H+)
551
O
724 (- CH3)
O
H2N
198
N
H
COOH
HN
623
539.2
683 (+ 2H+)
COOH
50
724.2
MH+ = 812
623.2
812.3
551.2
198.0
200
0
100
300
400
500
600
700
800
m/z (amu)
Figure 1. Extracted ion chromatogram (panel A) of an incubation mixture containing 1a (10 lM) and NADPH- and GSH-supplemented HLM. Panel B indicates the CID
spectrum of the MH+ ion (m/z 812) of the GSH conjugate at retention time = 14.45 min. The origins of the characteristic ions are as indicated. Both GSH conjugates exhibited
an identical mass spectrum suggesting that they were regioisomers. Assigned regiochemistry for GSH addition is arbitrary and for illustrative purposes only.

D. P. Walker et al. / Bioorg. Med. Chem. Lett. 18 (2008) 6071–6077
6077
2,4-dichloro-5-trifluoromethyl pyrimidine, see: Kath, J. C.; Richter, D. T.;
Luzzio, M. J. US Patent 7,122,670, October 16, 2006.
6. Compound 2k was prepared by
a
four step procedure according to the
following scheme:
CF3
N
CF3
N
a,b
PMB
N
H
N
N
H
Cl
N
Cl
NMe2
c,d
2k
Reagents and conditions: (a) 2.0 equiv ZnCl2 (1.0 M solution in ether),
CH2Cl2–tBuOH (1:1), À5 °C, 1 h; para-methoxybenzylamine, 1.1 equiv Et3N,
À5 °C ? room temperature 16 h, 25%; (b) (1R,2R)-N,N-dimethylcyclopentane-
1,2-diamine, Na2CO3, NMP, 80 °C, 2 h, 92%; (c) TFA–chloroform (5:1, 0.16 M),
50 °C, 16 h, 70%; (d) 4-bromopyridine, Pd2(dba)3 (0.02 equiv), Cs2CO3
(1.40 equiv), XANTPHOS (0.04 equiv), dioxane, 100 °C, 16 h, 52%.
7. Kath, J. C.; Luzzio, M. J. Patent Cooperation Treaty WO 2004/056807, July 08,
2004.
8. Avraham, H.; Park, S. Y.; Schinkmann, K.; Avraham, S. Cell. Signal. 2000, 12,
123.
9. Soglia, J. R.; Harriman, S. P.; Zhao, S.; Barberia, J.; Cole, M. J.; Boyd, J. G.; Contillo,
L. G. J. Pharm. Biomed. Anal. 2004, 36, 105.
10. Kalgutkar, A. S.; Fate, G.; Didiuk, M. T.; Bauman, J. Expert Rev. Clin. Pharmacol.
2008, 1, 515.
11. Glutathione-ethyl ester (GSH-EE) was used as the exogenous trapping agent in
the HTS reactive metabolite assay based on previous studies from our the
laboratories that revealed the improved sensitivity in sulfydryl conjugate
detection with ethyl ester derivative of GSH relative to the free carboxylic acid,
a feature attractive for HTS assays (see Ref. 9).
Figure 2. Crystal structure of compound 1n (magenta) bound to PYK2. Dashed lines
show hydrogen bond contacts to the hinge region.
12. Reactive metabolite assessment: HLM (P450 concentration = 0.3
containing test compound (10 M), GSH or GSH-EE (1 mM) and NADPH
(1.2 mM) in 100 mM potassium phosphate buffer (pH 7.4) were incubated for
30 min at 37 °C. The final incubation volume was 250 L. Samples without
either NADPH or test compound were used as negative controls. The reactions
were quenched by the addition of acetonitrile (375 L) and after centrifugation
lM)
l
we were able to transform a FAK-selective chemical series into
compounds with good PYK2 potency and 10- to 20-fold selectiv-
ity against FAK. It was later determined that the vast majority of
these compounds were positive in a HTS RM assay, indicating a
potential toxicological liability for this set of compounds. The 5-
aminooxindole moiety was identified as the likely culprit. Based
on the proposed mechanism for bioactivation, as well as a com-
bination of structure-based drug design and traditional medicinal
chemistry techniques, we prepared a follow-up series of PYK2
inhibitors that were negative in the RM assay. Furthermore,
many of these follow-up analogs and especially those possessing
the C4 cyclopentanediamine, showed good PYK2 activity and
FAK selectivity. Compounds 2c–2e, 2l, and 2n are further attrac-
tive in that they were stable in HLM.
l
l
(3000g, 15 min) the supernatant was concentrated under a steady nitrogen
stream. Sulfydryl conjugates were extracted from each sample using solid
phase extraction. A capillary liquid chromatography system (Dionex Corp.,
Sunnyvale, CA) was used for LC/MS/MS analysis. Chromatography was
performed in a VydacTM 300
lm id  5 cm C18 column that contained 5 lm
particles with a pore size of 300 Å (Grace Vydac, Hesperia, CA). Sulfydryl
conjugates were chromatographed using a binary mobile phase consisting of
acetonitrile:ammonium formate (5 mM):formic acid (10:90:0.05, v/v/v)
(solvent A) and acetonitrile:ammonium formate (5 mM):formic acid
(80:20:0.05, v/v/v) (solvent B) at a flow rate of 5 lL/min. A Thermo-Finnigan
Quantum triple quadrupole mass spectrometer with an orthogonal
electrospray ionization interface (Thermo-Finnigan Corp., San Jose, CA) was
used in the analysis. Electrospray ionization was initiated by applying voltage
of 3.8 kV (positive polarity). The auxiliary gas pressure and source transfer
capillary temperature were maintained at
throughout the study.
0 and 250 °C, respectively,
Acknowledgments
13. The fragment ion at m/z 198 was also observed in the CID spectrum of 1a.
Furthermore, metabolite identification study on 1a in NADPH-supplemented
human liver microsomes revealed the presence of the N-demethylated
metabolite (N-demethylation of the C4 sulfonamide group, MH+ = 493) with
a base fragment ion at m/z 184 (corresponding to loss of 14 amu from the
fragment ion at m/z 198).
We thank Jun Xiao, Kendra Nelson, Xumiao Zhao, and Samit
Bhattacharya for the preparation of compounds 1q, 2a, and 2b.
We thank Donn Wishka, Joel Arcari, Jim Southers, and Artie Brosius
for the preparation of compound 2j. We thank Elizabeth Schnippel
for the preparation of compounds 1b and 1j. We thank the ADME
CoE group for in vitro HTS ADME data.
H
CF3
N
N
O
N
H
N
NH
References and notes
R
O
O
N
S
198 (R = CH3)
184 (R= H)
1. Schlaepfer, D. D.; Hauck, C. R.; Sieg, D. J. Prog. Biophys. Mol. Biol. 1999, 71, 435.
2. (a) Guinamard, R.; Okigaki, M.; Schlessinger, J.; Ravetch, J. V. Nat. Immunol.
2000, 1, 36; (b) Okigaki, M.; Davis, C.; Falasca, M.; Harroch, S.; Felsenfeld, D. P.;
Sheetz, M. P.; Schlessinger, J. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 10740.
3. Buckbinder, L.; Crawford, D. T.; Qi, H.; Ke, H.-Z.; Olson, L. M.; Long, K. R.;
Bonnette, P. C.; Baumann, A. P.; Hambor, J. E.; Grasser, W. A.; Pan, L. C.; Owen, T.
A.; Luzzio, M. J.; Hulford, C. A.; Gebhard, D. F.; Paralkar, V. M.; Simmons, H. A.;
Kath, J. C.; Roberts, W. G.; Smock, S. L.; Guzman-Perez, A.; Brown, T. A.; Li, M.
Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 10619.
4. (a) Kalgutkar, A. S.; Gardner, I.; Obach, R. S.; Shaffer, C. L.; Callegari, E.; Henne,
K. R.; Mutlib, A. E.; Dalvie, D. K.; Lee, J. S.; Nakai, Y.; O’Donnell, J. P.; Boer, J.;
Harriman, S. P. Curr. Drug Metab. 2005, 6, 161; (b) Evans, D. C.; Watt, A. P.;
Nicoll-Griffith, D. A.; Baillie, T. A. Chem. Res. Toxicol. 2004, 17, 3; (c) Ju, C.;
Uetrecht, J. P. Curr. Drug Metab. 2002, 3, 367.
CH3
14. Baillie, T. A.; Davis, M. R. Biol. Mass Spectrom. 1993, 22, 319.
15. Meegala, S. K.; Wall, M. J.; Chen, J.; Wilson, K. J.; Ballentine, S. K.; DesJarlais, R.
L.; Schubert, C.; Crysler, C. S.; Chen, Y.; Molloy, C. J.; Chaikin, M. A.; Manthey, C.
L.; Player, M. R.; Tomczuk, B. E.; Illig, C. R. Bioorg. Med. Chem. Lett. 2008, 18,
3632.
16. Dahlin, D. C.; Miwa, G. T.; Lu, A. Y.; Nelson, S. D. Proc. Natl. Acad. Sci. U.S.A. 1984,
81, 1327.
17. The atomic coordinates for the X-ray structure of PYK2 in complex with
compound 1n have been deposited into the RCSB protein data bank: RCSB ID
Code rcsb049736, and under PDB ID Code 3ET7.
18. Koymans, L.; Donne-Op den Kelder, G. M.; te Koppele, J. M.; Vermeulen, N. P.
Xenobiotica 1993, 23, 633.
19. Slack-Davis, J. K.; Martin, K. H.; Tighman, R. W.; Iwanicki, M.; Ung, E. J.; Autry,
C.; Luzzio, M. J.; Cooper, B.; Kath, J. C.; Roberts, W. G.; Parsons, J. T. J. Biol. Chem.
2007, 282, 14845.
5. The addition of an aliphatic or aryl amine to 2,4-dichloro-5-
trifluoromethylpyrimidine is typically nonselective. This is in contrast to
many other 2,4-dichloro-5-substituted pyrimidines where the addition of an
amine is typically selective for the 4-position. The conditions for Method A were
specifically developed for the selective addition of amines to the 2-position of
20. Obach, R. S. Drug Metab. Dispos. 1999, 27, 1350.