Discovery of a Phosphoinositide 3-Kinase (PI3K) β/δ Inhibitor for the Treatment of Phosphatase and Tensin Homolog (PTEN) Deficient Tumors: Building PI3Kβ Potency in a PI3Kδ-Selective Template by Targeting Nonconserved Asp856

Article abstract of DOI:10.1021/acs.jmedchem.6b01821

Phosphoinositide 3-kinase (PI3K) β signaling is required to sustain cancer cell growth in which the tumor suppressor phosphatase and tensin homolog (PTEN) has been deactivated. This manuscript describes the discovery, optimization, and in vivo evaluation

Full text of DOI:10.1021/acs.jmedchem.6b01821

Subscriber access provided by UNIV OF CALIFORNIA SAN DIEGO LIBRARIES
Article
Discovery of a Phosphoinositide 3-Kinase (PI3K) b/d Inhibitor
for the Treatment of Phosphatase and Tensin Homolog
(PTEN) Deficient Tumors: Building PI3Kb Potency in a PI3Kd-
Selective Template by Targeting Non-Conserved Asp856
Stéphane Perreault, Jayaraman Chandrasekhar, Zhi-Hua Cui, Jerry B Evarts, Jia Hao, Joshua A. Kaplan,
Adam Kashishian, Kathleen S. Keegan, Thomas Kenney, David J Koditek, Latesh Lad, Eve-Irene Lepist, Mary
E. McGrath, Leena Patel, Bart Phillips, Joseph Therrien, Jennifer Treiberg, Anella Yahiaoui, and Gary Phillips
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b01821 • Publication Date (Web): 20 Jan 2017
Downloaded from http://pubs.acs.org on January 20, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 35
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Discovery of a Phosphoinositide 3-Kinase (PI3K) / Inhibitor for the Treatment of Phosphatase
and Tensin Homolog (PTEN) Deficient Tumors: Building PI3K Potency in a PI3K-Selective
Template by Targeting Non-Conserved Asp856
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Stephane Perreault,* Jayaraman Chandrasekhar, Zhi-Hua Cui, Jerry Evarts,† Jia Hao, Joshua A. Kaplan,
Adam Kashishian, Kathleen S. Keegan, Thomas Kenney, David Koditek, Latesh Lad, Eve-Irene Lepist,
Mary E. McGrath, Leena Patel, Bart Phillips, Joseph Therrien, Jennifer Treiberg, Anella Yahiaoui and
Gary Phillips.
Gilead Sciences, Inc., 199 E Blaine Street, Seattle, Washington 98102, United States
Gilead Sciences, Inc., 333 Lakeside Drive, Foster City, California 94404, United States
ABSTRACT
Phosphoinositide 3-kinase (PI3K)  signaling is required to sustain cancer cell growth in which the tumor
suppressor phosphatase and tensin homolog (PTEN) has been deactivated. This manuscript describes the
discovery, optimization, and in vivo evaluation of a novel series of PI3K inhibitors in which PI3K
potency was built in a PI3K-selective template. This work led to the discovery of a highly selective
PI3K/ inhibitor displaying excellent pharmacokinetic profile and efficacy in a human PTEN-deficient
LNCaP prostate carcinoma xenograft tumor model.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 35
1
2
3
4
INTRODUCTION
5
6
7
8
9
The phosphoinositide 3-kinase (PI3K) pathway plays a role in a wide range of key cellular functions such
as cell growth, proliferation and survival.1,2,3,4 Aberrant activation of the PI3K/AKT pathway has been
observed with high frequency in human malignancies, suggesting a significant role in tumorigenesis,
cancer progression and treatment resistance.5,6 Class I PI3Ks are heterodimers bearing regulatory and
catalytic subunits, which are further subdivided into class IA and IB.7 Class IA is comprised of three
isoforms of the catalytic subunit: p110, p110 and p110, while p110 is the only isoform in class IB.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
PI3K and  are ubiquitously expressed in all tissues, while PI3K and  are confined to leukocytes.
Class I PI3Ks catalyze the phosphorylation of phosphatidylinositol-4,5-bisphosphate (PIP2) to produce
the second messenger phosphatidylinositol-3,4,5-trisphosphate (PIP3).8 Pan inhibitors of the Class I
PI3Ks, along with molecules that inhibit additional downstream targets, such as mTOR, have provided
evidence that attenuating signaling through this node can yield clinical benefit in oncology.9 However, the
utility of pan inhibitors has been limited by side effects, such as hyperglycemia, resulting in dose
reductions that erode efficacy. Most recent efforts are devoted to target the relevant oncogenic PI3K
isoform with isoform-selective inhibitors in order to circumvent off-target toxicities and to provide a
larger therapeutic window.10,11,12,13,14,15,16
In many solid tumors associated with aberrant signaling of the PI3K isoform, the increased signaling
through the PI3K pathway is not induced via activating mutations in PI3K, but rather by deactivation of
the tumor suppressor phosphatase and tensin homolog (PTEN).17,18 The protein encoded by this gene is a
phosphatidylinositol-3,4,5-trisphosphate 3-phosphatase which functions as a negative regulator of the
lipid kinase activity. Preclinical studies have shown that down-regulation of PI3K in PTEN-deficient
cancer cells results in pathway inactivation and subsequent inhibition of growth in both cell-based and in
vivo settings.17 This discovery provided strong rationale for the development of PI3K-selective
ACS Paragon Plus Environment

Page 3 of 35
Journal of Medicinal Chemistry
1
2
3
4
inhibitors. GSK-263677110,19 and AZD818610,20 are -weighted / inhibitors currently under clinical
5
6
evaluation for the treatment of castration-resistant prostate cancer (Figure 1).21,22,23,24,25,26,27
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. PI3K inhibitors in clinical trials.
In this manuscript, we describe our efforts toward identifying potent, selective, and metabolically stable
PI3K inhibitors suitable for the treatment of PTEN-deficient tumors. Our goals were threefold: potent
inhibition of PI3K in PTEN-deficient PC3 prostate cancer cells (inhibition of pAKT with EC50 < 20
nM), isoform selectivity of greater than 100-fold over PI3K and PI3K in order to allow EC90 coverage
of PI3K at trough concentrations without compromising isoform selectivity, and in vitro human
predicted clearance of 0.2 L/h/kg or below to allow for a twice-daily dosing regimen.
RESULTS AND DISCUSSION
In an effort to develop isoform-selective PI3K inhibitors, we envisaged building PI3K potency from our
previously disclosed PI3K-selective template, exemplified by idelalisib (1).28,29 As revealed by its X-ray
crystal structure in the PI3Kδ isoform, the quinazolinone of idelalisib occupies a specificity pocket
between Met752 and Trp760, a closed pocket in the apo structure of the enzyme (Figure 2a).30 This
binding mode may imbue propeller shaped inhibitors such as idelalisib with improved selectivity over
other lipid kinases that show reduced conformational flexibility in this region.31 This orthogonal
arrangement orients the phenyl group to sit on the hydrophobic region II (H2). While many residues in
the ATP pocket are conserved among all four class I PI3K isoforms, sequence alignment of the residues
in the hydrophobic region II shows several variations, such as Asp856, which is unique to PI3K (Figure
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 35
1
2
2b).32 Accordingly, we reasoned that a well-placed hydrogen bond donor could engage in an interaction
with Asp856 resulting in an improvement in PI3K potency.33 Proof of concept was achieved through the
introduction of a 4-hydroxy group on the phenyl ring of idelalisib (Table 1, 1 vs. 2). This modification
leads to a 1000-fold improvement in PI3K potency. With our knowledge that the primary site of
metabolism of idelalisib is the purine moiety,34,35 we also evaluated our hypothesis in a second analogous
series bearing a 4-amino-5-cyanopyrimidine hinge binder and a similar increase in PI3K potency was
observed (3 vs. 4). From these results, we began a medicinal chemistry effort toward the identification of
potent and selective PI3Kβ/δ inhibitors.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. a) 2.4 Å X-ray crystal structure of idelalisib (1) bound in the PI3K ATP binding site (pdb: 4XE0; coordinates
available).30 Asp832 and Asn836 (highlighted in yellow) differ among the PI3K isoforms. For clarity, all other residues have
been removed. Blue dashed lines show hydrogen bond contacts between the inhibitor and the protein. b) Different residues in the
hydrophobic region II among class I PI3Ks. Asn836 of PI3K corresponds to Asp856 of PI3K.
ACS Paragon Plus Environment

Page 5 of 35
Journal of Medicinal Chemistry
1
2
3
4
Table 1. Building PI3K potency in a PI3K-selective template
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1 (idelalisib)
7500
2
1600 (593x)
2.7
3
4
5
82 (33x)
2.5
6
3490 (249x)
14
PI3KIC50 (nM) a,b
PI3KIC50 (nM) a
PI3KIC50 (nM) a
2025
54
281 (400x)
0.7
3700
18
14
0.34
0.3
0.6
3.2
PI3KIC50 (nM) a,b

2100
4300 (1592x)
154
187 (267x)
990 (396x)
990 (71x)
a The activity against each class I PI3K was evaluated in in vitro kinase assays containing 2xKm steady state concentrations of ATP (average of ≥ 2 determinations). b
Numbers in parentheses represent fold selectivity over PI3K (PI3K or PI3K IC50 divided by PI3K IC50).
Phenol-containing compounds are prone to pharmacokinetic (fast glucuronidation as prelude to excretion)
and/or toxicological limitations (formation of quinones as chemically reactive metabolites). To eliminate
the phenol, but retain the proposed hydrogen bond interaction between the inhibitor and Asp856, a variety
of moieties were prepared and tested, including, but not limited to, heterocycles, such as indoles,
indazoles, benzotriazoles, pyrazoles, and acylated or sulfonylated (hetero)aryl amines such as carbamates
and ureas.36,37,38,39 Our studies indicated that pyrazoles 5 (pyrazol-3-yl) and 6 (pyrazol-4-yl) appeared to
be attractive starting points (Table 1). Further profiling of compounds 5 and 6 revealed very high
predicted clearance in human hepatocytes (>0.95 L/hr/kg). This observation is consistent with our
previously disclosed PI3K-selective series: the 4-amino-5-cyanopyrimidine hinge binder is a substrate
for aldehyde oxidase-mediated metabolism leading to C2-oxidation on the pyrimidine.13,40 Despite the
lower selectivity for the PI3K isoform, we first decided to investigate the pyrazol-3-yl series due to the
favorable PI3K potency of 5.
Alternative substitutions around the pyrimidine hinge binder were explored in an effort to mitigate the
high predicted clearance of 5 (Table 2). As discovered previously, introduction of a primary C2-amine is
well tolerated and prevents oxidation of the pyrimidine (7).12-13 The low predicted hepatic clearance was
attractive but dampened by the concomitant increase in PI3K and PI3K potency. It is likely that the
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 35
1
2
3
4
5
6
7
8
increased potency in all isoforms is the result of a third hydrogen bond involving the 2-amino moiety of
the pyrimidine hinge binder. Additionally, compound 7, as well as closely related cyclopropyl analog 8,
displays very low forward permeability as determined by a Caco-2 cell monolayer assay.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Our first approach to improving the permeability of this series consisted of reducing the total polar
surface area (tPSA) and increasing the lipophilicity as measured by cLogP. Substitution of the nitrile
group at the C5-position of the pyrimidine by a hydrogen atom (9) contributes to a reduction in tPSA (170
Å2 to 146 Å2). Unfortunately, this modification results in a substantial loss in potency. Replacement of the
nitrile group by a chlorine atom (10) is well tolerated in terms of PI3K potency. This change is also
accompanied by a desirable increase in forward permeability, presumably the result of a higher cLogP
and lower tPSA (146 Å2). A more lipophilic cyclopropyl group at R1 (11) leads to an even greater
improvement in permeability. Moreover, compounds 10 and 11 afford better selectivity over the PI3K
and PI3Kisoforms relative to the nitrile analogs 7 and 8. The predicted clearances of 10 and 11 are
higher but remain in an acceptable range for further consideration. Metabolite identification studies on
compound 11, after incubation with human hepatocytes, reveal the formation of an oxidative metabolite
(M+16) of the pyrimidine moiety. This metabolic pathway is not observed in compounds with the
electron withdrawing nitrile group at the C5-position of the pyrimidine (7 and 8).
As a second approach to improve the permeability, we envisaged reducing the number of hydrogen bond
donors. Replacement of the primary amine at R3 by a methyl group (12) or a chlorine atom (13) results in
an increase in forward permeability at the expense of higher predicted clearances. The loss in metabolic
stability can be rescued by introducing a difluoromethyl group at R3 (14). This modification is also
accompanied by a substantial improvement in forward permeability. The trifluoromethyl analog 15
exhibits low PI3K potency. We believe this is the result of an inevitable electrostatic repulsion between
the third fluorine and the carbonyl group of Glu846 in PI3K (vide infra).
ACS Paragon Plus Environment

Page 7 of 35
Journal of Medicinal Chemistry
1
2
3
4
Table 2. Optimization of the 2-aminopyrimidine hinge binder
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
IC50 (nM)a
CLpr
Caco-2
Compound
R1
R2
R3
cLogP
(L/hr/kg)c
(AB/BA)d
PI3Kb
23 (33x)
PI3K
0.7
PI3K
0.2
PI3Kb
54 (77x)
7
CH3
CN
CN
H
NH2
NH2
NH2
NH2
NH2
CH3
Cl
1.48
1.92
1.93
2.70
3.14
2.57
2.54
2.27
2.40
0.06
0.05
ND
0.05/10
0.16/14
ND
8
89 (56x)
1.6
0.6
199 (124x)
>10000 (>6x)
1247 (594x)
4150 (525x)
267 (157x)
1300 (250x)
1131 (122x)
>10000 (>15x)
9
CH3
CH3
>10000 (>6x)
302 (144x)
1074 (136x)
45 (26x)
1584
2.1
317
2.1
10
11
12
13
14
15
Cl
0.22
0.20
0.40
0.73
0.16
ND
1.1 / 20
2.2/12
2.0/17
1.3/24
9.0/42
ND
Cl
7.9
5.2
CN
CN
CN
CN
1.7
0.8
308 (59x)
5.2
5.9
CHF2
CF3
686 (74x)
9.3
13
CH3
>10000 (>15x)
671
309
a The activity against each class I PI3K was evaluated in in vitro kinase assays containing 2xKm steady state concentrations of ATP (average of ≥ 2 determinations). b
Numbers in parentheses represent fold selectivity over PI3K (PI3K or PI3K IC50 divided by PI3K IC50). c Determined from human hepatocytes. d 106 cm/s. ND =
Value not determined.
These results prompted us to further investigate different quinazolinone and R1 substitutions in the 2,6-
diamino-5-chloropyrimidine and the 2-amino-5-cyano-4-(difluoromethyl)pyrimidine hinge binder series.
While exploring alternate substitutions on the quinazolinone ring, we found that minor modifications
have substantial impact on potency and selectivity (Table 3). Substituents such as 5-fluoro (16a-b), 6-
fluoro (17a-b), and 8-fluoro (18a-b) all result in a loss of PI3K potency. Interestingly, the 6,8-
difluoroquinazolinone motif found in 19a-b results in a 2-fold increase in PI3K potency relative to the
monofluorinated analogs 17a-b and 18a-b. Potency and isoform selectivity can be further improved by
replacing the 8-fluoro with an 8-chloro (20a-b). Combining the 8-chloro-6-fluoroquinazolinone with
various R1 moieties (21a-b and 22a-b) results in a loss of potency. Nonetheless, the biochemical profile
of compound 21b remains promising based on potency and isoform selectivity.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 35
1
2
Table 3. SAR results for quinazolinone ring and R1 modifications
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
IC50 (nM)a
IC50 (nM)a
Core
Compound
16a
Compound
16b
PI3Kb
PI3K
PI3K
PI3Kb
PI3Kb
PI3K
PI3K
PI3Kb
1085
(31x)
>10000
(>286x)
1511
(14x)
>10000
(>92x)
35
21
109
121
29
51
322
(8x)
>10000
(>238x)
720
(6x)
7989
(66x)
17a
18a
19a
20a
21a
42
34
17
7.8
28
9.8
12
17b
18b
19b
20b
21b
26
9.5
6.2
3.3
7.2
357
(11x)
>10000
(>294x)
187
(6x)
4262
(147x)
291
(17x)
4407
(259x)
171
(11x)
2311
(154x)
3.0
5.3
15
15
850
(109x)
>10000
(>1282x)
284
(109x)
1277
(491x)
2.6
4.5
2080
(74x)
>10000
(>357x)
756
(168x)
3466
(770x)
6937
(50x)
>10000
(>72x)
1303
(81x)
7836
(490x)
22a
138
132
22b
16
13
a The activity against each class I PI3K was evaluated in in vitro kinase assays containing 2xKm steady state concentrations of ATP (average of ≥ 2 determinations). b
Numbers in parentheses represent fold selectivity over PI3K (PI3K or PI3K IC50 divided by PI3K IC50).
The improvement in PI3K potency observed with the introduction of the C2-amine on the pyrimidine
hinge binder (5 vs. 7) led us to further explore the pyrazol-4-yl series exemplified by compound 6 (Table
1). However, in a head-to-head comparison between the pyrazol-3-yl series (23) and the corresponding
pyrazol-4-yl series (24) (SAR pair analysis of compounds in which the only variable is the connectivity
point on the pyrazole, Figure 3a), we observed that the pyrazol-4-yl leads to inferior PI3Kpotency in
most cases (Figure 3b). In terms of selectivity, it appears that the combination of the pyrazol-3-yl with the
2,6-diamino-5-chloropyrimidine (green circles) leads to greater PI3K selectivity (Figure 3c). In a third
analysis, we observed that the pyrazol-3-yl analog generally leads to greater PI3K selectivity when
ACS Paragon Plus Environment

Page 9 of 35
Journal of Medicinal Chemistry
1
2
3
4
5
6
compared to the corresponding pyrazol-4-yl compound (Figure 3d). This last trend was independent of
the hinge binder and led us to abandon the pyrazol-4-yl series.
7
8
9
a) SAR pair analysis: pyrazol-3-yl vs. pyrazol-4-yl
b) PI3K potency: pyrazol-3-yl vs. pyrazol-4-yl
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
N
NH
O
N
O
N
Xn
Xn
NH
N
N
N
R1
R1
HN
N
NH2
HN
N
NH2
N
N
23
pyrazol-3-yl
24
pyrazol-4-yl
R2
R2
R3
R3
R2 = CN and R3 = NH2
R2 = Cl and R3 = NH2
R2 = CN and R3 = CHF2
PI3K IC50 of pyrazol-4-yl (nM)
c) PI3K / PI3K selectivity: pyrazol-3-yl vs. pyrazol-4-yl
d) PI3K / PI3K selectivity: pyrazol-3-yl vs. pyrazol-4-yl
PI3K IC50 / PI3K IC50 PI3K of pyrazol-4-yl
PI3K IC50 / PI3K IC50 of pyrazol-4-yl
Figure 3. SAR pair analysis illustrating that the pyrazol-3-yl series (23) is superior to the pyrazol-4-yl series (24) in terms of
PI3K potency and PI3K selectivity. Each colored shape represents a pair of compounds 23 and 24 with identical substituents
Xn, R1, R2, and R3, the only variable being the connectivity point on the pyrazole (pyrazol-3-yl vs. pyrazol-4-yl). The value on the
x-axis corresponds to the pyrazol-4-yl compound 24 and the value on the y-axis is for the corresponding pyrazol-3-yl compound
23.
The binding mode of this series of inhibitors has not been characterized crystallographically in PI3K.41
However, it can be modeled in a PI3K homology model (coordinates have been included in Supporting
Information) based on the crystal structure of PI3Kδ co-crystallized with recently disclosed (S)-2,4-
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 35
1
2
3
4
diamino-6-((5-chloro-8-fluoro-4-oxo-3-(pyridin-3-yl)-3,4-dihydroquinazolin-2-
5
6
7
8
yl)(cyclopropyl)methylamino)-pyrimidine-5-carbonitrile (GS-9901, pdb: 5T8I).12 Compounds such as 11
and 14 are likely to bind to the ATP binding site of the kinase domain of PI3Kβ (Figure 4). Similar to
other propeller shaped PI3K inhibitors, the quinazolinone occupies an induced specificity pocket between
Met773 and Trp781 of PI3Kβ. Additionally, molecular modeling indicates that the hydrogen bond donor
in the pyrazole ring can form an energetically favorable interaction with Asp856. The 2,6-diamino-5-
chloropyrimidine of 11 is expected to serve as a three-point hinge binder (Figure 4a). On the other hand,
the 2-amino-5-cyano-4-(difluoromethyl)pyrimidine moiety of compound 14 is likely to bind to the hinge
with two hydrogen bonds (Figure 4b). Modeling also suggests a weakly attractive interaction between the
hydrogen of the 4-difluoromethyl group and the carbonyl of Glu846. Previously discussed compound 15,
bearing a 4-trifluoromethyl group on the pyrimidine, most likely suffers from a repulsive interaction with
the same Glu846 carbonyl.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. Docked poses of compound 11 (a) and 14 (b) in the homology model of PI3K. In both cases, the pyrazole-NH is
predicted to hydrogen bond to Asp856. For clarity, all other residues have been removed. Blue dashed lines show predicted
hydrogen bond contacts between the inhibitor and the protein. The PI3Kβ homology model used was built from the crystal
structure pdb: 5T8I (coordinates have been included in Supporting Information).12
ACS Paragon Plus Environment

Page 11 of 35
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
To shed some light on this possible interaction between the pyrazole moiety and Asp856, we synthesized
the corresponding isoxazole 25 (Table 4). The ring lacks the hydrogen bond donor and leads to a 250-fold
drop in PI3K potency while PI3K remains equipotent. This result supports the prediction of an
interaction between the hydrogen bond donating pyrazole and Asp856 in this series of PI3K inhibitors.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 4. Pyrazol-3-yl vs. isoxazol-3-yl
10
302
2.1
25
3538
539
1.9
PI3KIC50 (nM) a
PI3KIC50 (nM) a
PI3KIC50 (nM) a
PI3KIC50 (nM) a
2.1
1247
499
a The activity against each class I PI3K was evaluated in in vitro kinase assays containing 2xKm steady state concentrations of ATP (average of ≥ 2 determinations).
Compounds meeting potency and selectivity criteria were then evaluated for tumor cell activity by
measuring inhibition of AKT phosphorylation in the PTEN-deficient PC3 prostate cancer cell line (Table
5). Correlation between biochemical and cellular potency is more consistent for compounds with the 2,6-
diamino-5-chloropyrimidine hinge binder (10, 11 and 20a), while larger shifts were observed with the 2-
amino-5-cyano-4-(difluoromethyl)pyrimidine (14, 20b and 21b). This observation, combined with higher
predicted human clearances for compounds 20b and 21b, led us to focus on analogs with the 2,6-diamino-
5-chloropyrimidine hinge binder (10, 11 and 20a).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 35
1
2
3
4
Table 5. Activity against AKT phosphorylation in PC3 cells and predicted human hepatic clearance
5
6
7
8
PC3 EC50
(nM)a
CLpr
PI3K IC50
Compound
(L/hr/kg)b
(nM)a
10
11
2.1
7.9
7.8
9.3
2.6
4.5
4.2
7.6
20
45
15
42
0.22
0.20
0.17
0.16
0.29
0.34
9
20a
14
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
20b
21b
a Average of ≥ 2 determinations. b Determined from human hepatocytes.
We next evaluated the pharmacokinetic profiles of the most promising compounds 10, 11 and 20a.
Despite moderate oral bioavailability, the in vivo rat pharmacokinetic profile of 10 was promising,
exhibiting good correlation between in vitro and in vivo clearances (Table 6). Introduction of a more
lipophilic cyclopropyl at R1 (11) enhances the forward permeability leading to an overall improvement in
oral bioavailability. Compound 20a, bearing an 8-chloro-6-fluoroquinazolinone, leads to an oral
bioavailability of 66% in rat. We attribute this result to lower observed clearance and improved forward
permeability as measured by a Caco-2 cell monolayer assay.
Inspired by the rat pharmacokinetic profiles of 11 and 20a, we envisaged combining the positive features
of these two analogs in one compound along with the metabolically stable 2,6-diamino-5-
cyanopyrimidine hinge binder exemplified in compounds 7 and 8. The resulting compound (26) yields a
promising amalgamation of both bio- and physicochemical properties including a significant
improvement in forward permeability in comparison to 7 and 8. Encouraged by the low predicted
clearance and the improved permeability, we progressed compound 26 to in vivo pharmacokinetic
evaluation. In addition to poor correlation between predicted and observed clearances, oral dosing of 26 at
5 mg/kg in rat suffers from a disappointing bioavailability of 12%. Among these four compounds, 20a is
the most attractive based on its combination of potency and selectivity, lower predicted and observed
clearances, and bioavailability.
ACS Paragon Plus Environment

Page 13 of 35
Journal of Medicinal Chemistry
1
2
3
4
Table 6. Summary of pharmacokinetic profiles in rat
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
10
302 / 2.1 / 2.1 / 1248
4.2
11
1074 / 7.9 / 5.2 / 4150
7.6
20a
850 / 7.8 / 5.3 / >10000
20
26
84 / 0.7 / 0.6 / 632
3.7
PI3K /IC50 (nM) a
EC50 PC3 (nM) a
Caco-2 (AB / BA)b
CLpr (h) (L/h/kg) c
CLpr (r) (L/h/kg) c
CLobs (r) (L/h/kg) d
Vss (r) (L/kg) d,e
1.1 / 19.6
0.22
2.2 / 11.8
0.20
5.7 / 21.7
0.17
1.5 / 17.3
0.05
1.45
1.37
0.88
0.81
1.43 ± 0.26
1.29 ± 0.10
26 ± 12
0.58 ± 0.07
0.83 ± 0.10
42 ± 22
0.26 ± 0.02
0.52 ± 0.03
66 ± 17
0.07 ± 0.00
0.22 ± 0.01
12 ± 5
F (r) (%)f,g
a Average of ≥ 2 determinations. b 106 cm/s. c Predicted clearance in hepatocytes. d Intravenous dose of 1 mg/kg. e Volume of distribution at steady state. f Oral dose of
5 mg/kg. g Number of animals per dosing route = 3. ND = Value not determined. h = human, r = rat.
A summary of the pharmacokinetic parameters in additional preclinical species after intravenous and oral
administration of 20a is presented in Table 7. The in vivo data indicate that 20a has low to intermediate
total clearance (CL) in comparison to hepatic blood flow in all species evaluated. Volumes of distribution
(Vss) are higher than total body water in all species with the exception of rat in which higher plasma
protein binding (PPB) is observed. The apparent elimination half-life (t1/2) and mean residence time
(MRT) are 2 to 3 hours in rat, 4 to 5 hours in dog and rhesus monkey, and 5 to 6 hours in cynomolgus
monkey. More than half of the administered dose is orally bioavailable in rat, dog and cynomolgus
monkey. The high oral bioavailability is consistent with low observed clearance.
Further characterization of 20a revealed that it does not inhibit major human CYP450 enzymes (IC50 > 25
uM) or hERG (IC50 > 25 M) in binding assays. Compound 20a does not significantly interact with any
protein kinases using the KINOMEscan™ (DiscoveRx) platform at a compound concentration of 10 M,
validating that 20a is a potent and selective inhibitor of PI3K/δ (see Supporting Information).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 35
1
2
3
4
Table 7. Summary of pharmacokinetic parameters for 20a in preclinical species
5
6
7
8
Sprague-
Dawley rat
Cynomolgus
monkey
Rhesus
monkey
Parameter
PPB (% free)a
3H CLpr (L/h/kg) b
CL (L/h/kg) c
Vss (L/kg) d
Beagle dog
1.96
0.85
3.05
3.33
0.89
0.67
0.25
0.58
9
0.26 ± 0.02
0.52 ± 0.03
2.69 ± 0.27
1.97 ± 0.20
66.4 ± 17.5
0.25 ± 0.11
1.14 ± 0.36
3.91 ± 0.23
4.77 ± 0.56
90.4 ± 21.7
0.24 ± 0.07
1.50 ± 0.34
4.96 ± 0.46
6.28 ± 0.59
68.6 ± 10.5
0.41 ± 0.03
1.88 ± 0.17
3.61 ± 0.21
4.54 ± 0.19
ND
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Terminal t 1/2 (h) e
MRT (h) f
F (%)g
a Plasma protein binding. b Predicted clearance in hepatocytes using 3H compound. Intravenous dose of 1 mg/kg. d Volume of distribution at steady state. e Half-life
c
iv. f Mean residence time iv. g Oral doses of 20a (5 mg/kg) were formulated in 5% ethanol, 50% PEG 300, and 45% water. ND = Value not determined. (Mean ± SD,
n=3)
As previously discussed, in PTEN-deficient tumors, the counterbalance to constitutive PI3K activity is
absent, leading to increased signaling through the PI3K/AKT pathway. The phosphorylation state of AKT
has been extensively used as a downstream measure of PI3K activity.12-27 Downstream signaling effects
of 20a inhibition have been characterized in a subset of PTEN-deficient tumor cell lines that show
dependence on PI3K for viability: the prostate carcinomas PC3 and LNCaP, and the breast
adenocarcinomas MDA-MB-415 and ZR-75-1 (Table 8). We also included the kidney carcinoma A498
cell line which has a PI3K activating mutation. Compound 20a inhibits the phosphorylation of AKT1
Ser473 in these PI3K-dependent cell lines with EC50 values ranging from 3.4 nM to 7.3 nM. Cell
viability was also found to be potently inhibited by 20a with GI50 ranging from 2 to 120 nM.
Table 8. Activity of 20a against AKT1 Ser473 phosphorylation in PI3K-dependent tumor cell lines
Cell line
PC3a
EC50 (nM)c
GI50 (nM)
5.4
7.3
5.7
3.4
5.6
120
38
12
2
LNCaPa
MDA-MB-415a
ZR-75-1a
A498b
80
a PTEN-deficient. b PI3K mutant. c Data determined from the ratio of Units pAKT1 Ser473 per nanogram total AKT1.
The activity of 20a was assessed in a human PTEN-deficient, androgen-responsive LNCaP xenograft
tumor model. As shown previously, AKT Ser473 phosphorylation and cell viability of LNCaP tumor cells
are potently inhibited by PI3Kβ inhibitor 20a (EC50 = 7.3 nM and GI50 = 38 nM). Mice engrafted with
LNCaP tumor cells were treated with 20a at 15 mg/kg, 7.5 mg/kg, and 3 mg/kg twice a day for 13
ACS Paragon Plus Environment

Page 15 of 35
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
consecutive days following the establishment of the tumor (Figure 5). Tumor burden expressed as volume
(mm3) was measured twice a week throughout the experiment. After the final dose, treatment with 20a
resulted in tumor growth inhibition of 73%, 65% and 42% in the 15 mg/kg, 7.5 mg/kg and 3 mg/kg dose
groups, respectively (p<0.0001). There was no statistical difference between the tumor growth inhibition
at 7.5 mg/kg and 15 mg/kg suggesting a maximal efficacious dose of 7.5 mg/kg in this experiment.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. Activity of 20a in a human PTEN-deficient, androgen-responsive LNCaP xenograft tumor model in mice. 20a was
dosed twice a day for 13 consecutive days following the establishment of the tumor.
Inhibition of pAKT1 Ser473 was assessed in the LNCaP xenograft model at study completion. After the
final dose, each treatment group was divided into subgroups of 3 mice. Plasma and tumors were harvested
at 2, 8 or 12 hours post dose. Plasma levels of 20a were determined and tumors were lysed and assayed
for the presence of pAKT1 and total AKT1 at each time point (Figure 6). Maximum inhibition of AKT1
phosphorylation (64%) was observed at 8 hours in the 7.5 mg/kg and 2 hours in the 15 mg/kg treatment
groups. Target coverage in the tumor was compared to the protein adjusted EC90 (highest dotted line in
Figure 6). At the 12 hour time point, coverage over protein adjusted EC90 was shown to be 7x and 17x for
the 3 mg/kg and 7.5 mg/kg doses, respectively.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 35
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 6. AKT1 phosphorylation levels in LNCaP xenograft tumors treated with 20a with corresponding mice plasma levels.
Data presented as the ratio of Units pAKT1 Ser473 per nanogram of total AKT1. The lower and higher dashed lines represent
protein adjusted EC50 and EC90, respectively.
CHEMISTRY
Analogs 1-26 were synthesized following the previously described general synthetic route shown in
Scheme 1.12-13 The quinazolinone core 31 is constructed via a one-pot, 2-step dehydrative cyclization.42
Oxazinone intermediate 29 is prepared by treatment of an anthranilic acid 27 with a protected amino acid
28 in the presence of diphenyl phosphite in pyridine. Subsequent addition of aniline (30a), 4-((tert-
butyldimethylsilyl)oxy)aniline (30b), 3-amino-1H-pyrazole (30c), 4-amino-1H-pyrazole (30d) or 3-
aminoisoxazole (30e) to the reaction mixture results in the formation of the quinazolinone 31. Removal of
the tert-butoxycarbonyl group of 31 (as well as the tert-butyldimethylsilyl group in the case of the 31c
and 31d) with trifluoroacetic acid affords the primary amine 32 which is reacted with a substituted
halopyrimidine 33 for the incorporation of the hinge binder to afford the final PI3K/ inhibitors (1-26).
ACS Paragon Plus Environment

Page 17 of 35
Journal of Medicinal Chemistry
1
2
Scheme 1. General synthetic sequencea
3
4
5
6
7
8
9
O
N
O
N
O
N
Xn
6
Xn
Xn
6
5
8
5
8
O
Ar
O
Ar
Xn
5
6
3
N
N
O
i
30a-e
33a-i
R1
R1
NHR5
OH
+
HO
R1
R1
7
7
ii
iv
4
NHBoc
NH2
27a-f
HN
R2
N
R4
NHBoc
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
31a-o (R5 = Boc)
32a-o (R5 = H)
28a-c
29
N
iii
28a R1 = CH2CH3
28b R1 = CH3
28c R1 = cyclopropyl
27a Xn = 6-F
27b Xn = 6-Cl
27c Xn = 5-F
27d Xn = 3-F
27e Xn = 3,5-F2
27f Xn = 3-Cl-5-F
1-26
R3
31a-32a Xn = 5-F; R1 = CH2CH3; Ar = C6H5
31b
Xn = 5-F; R1 = CH2CH3; Ar = 4-(OTBS)C6H4
Xn = 5-F; R1 = CH2CH3; Ar = 4-(OH)C6H4
32b
31c-32c Xn = 5-Cl; R1 = CH3; Ar = C6H5
OTBS
n = 5-Cl; R1 = CH3; Ar = 4-(OTBS)C6H4
Xn = 5-Cl; R1 = CH3; Ar = 4-(OH)C6H4
N
31d
X
NH
NH
O
32d
N
N
H2N
H2N
H2N
H2N
H2N
31e-32e Xn = 5-Cl; R1 = CH3; Ar = pyrazol-3-yl
31f-32f Xn = 5-Cl; R1 = CH3; Ar = pyrazol-4-yl
31g-32g Xn = 5-Cl; R1 = cyclopropyl; Ar = pyrazol-3-yl
31h-32h Xn = 5-F; R1 = CH3; Ar = pyrazol-3-yl
31i-32i Xn = 6-F; R1 = CH3; Ar = pyrazol-3-yl
31j-32j Xn = 8-F; R1 = CH3; Ar = pyrazol-3-yl
31k-32k Xn = 6,8-F2; R1 = CH3; Ar = pyrazol-3-yl
31l-32l Xn = 8-Cl-6-F; R1 = CH3; Ar = pyrazol-3-yl
30a
30b
30c
30d
30e
33a 6-chloro-9-(tetrahydro-2H-pyran-2-yl)-9H-purine
33b X = Cl; R2 = CN; R3 = NH2; R4 = H
33c X = Cl; R2 = CN; R3 = NH2; R4 = NH2
33d X = Cl; R2 = H; R3 = NH2; R4 = NH2
33e X = F; R2 = Cl; R3 = NH2; R4 = NH2
33f X = Cl; R2 = CN; R3 = CH3; R4 = NH2
33g X = Cl; R2 = CN; R3 = Cl; R4 = NH2
33h X = Cl; R2 = CN; R3 = CHF2; R4 = NH2
33i X = Cl; R2 = CN; R3 = CF3; R4 = NH2
X
N
R4
N
R2
R3
31m-32m Xn = 8-Cl-6-F; R1 = cyclopropyl; Ar = pyrazol-3-yl
31n-32n Xn = 8-Cl-6-F; R1 = CH2CH3; Ar = pyrazol-3-yl
31o-32o Xn = 5-Cl; R1 = CH3; Ar = isoxazol-3-yl
33a-i
a Reagents and conditions: (i) Pyridine, diphenyl phosphite, 40 C, 2 h; (ii) 50 C 3-20 h, 50-100%; (iii) CF3CO2H, CH2Cl2, rt,
0.5-4 h, quant.; (iv) i-Pr2NEt, isopropanol, 80-130 C W or thermal 80-130 C, 1-24 h, 30-90%. Boc = tert-butoxycarbonyl, TBS
= tert-butyldimethylsilyl.
CONCLUSIONS
In summary, we have discovered and optimized a series of potent and highly selective PI3K/δ inhibitors.
This was achieved by building PI3K potency in a PI3K-selective template by introducing a pyrazole
moiety which presumably engages an interaction with Asp856. The pyrimidine hinge binder was
optimized to yield adequate metabolic stability and forward permeability. Modification of the
quinazolinone substituents led to improved isoform selectivity, culminating in 20a, a potent and selective
inhibitor of PI3K/δ with good pharmacokinetic properties. Moreover, 20a demonstrated efficacy in a
human PTEN-deficient LNCaP prostate carcinoma xenograft tumor model. Additional studies will be
presented in future publications.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 35
1
2
3
4
EXPERIMENTAL SECTION
5
6
7
8
9
General Procedures. All commercial reagents and anhydrous solvents were used as provided without further
purification. Microwave reactions were completed using a benchtop CEM microwave reactor. Flash
chromatography was performed using an ISCO Combiflash or a Biotage Isolera purification system with Silicycle
pre-packed silica gel cartridges. Final compounds were purified by prep HPLC on a Gilson 271 system, with a
Phenomenex Kinetex C18 Column (250 mm x 30 mm, 100 A, 5  and a gradient elution of acetonitrile/water with,
5-95% for 20 mins at a flow rate of 35 mL/min. For all samples, 0.1% TFA was added to both eluents. All 1H NMR
spectra were recorded on a Varian 400-MR 400 MHz spectrometer. Proton chemical shifts are reported in parts per
million from an internal standard or residual solvent. The purity of the tested compounds was assessed by an Agilent
LCMS. A Silicycle Siliachrom SB C18 column (4.6 x 50 mm, 3 ) was used with a gradient elution of acetonitrile in
water, 0−95% for 7 min at a flow rate of 0.9 mL/min with detection at 256 nm. Alternatively, a Chromolith column
(50 x 2 mm) was used with a gradient elution of acetonitrile in water, 0−95% for 3 min at a flow rate of 1.9 mL/min
with detection at 256 nm wavelength. For all samples 0.1% TFA was added to both eluents. All final compounds
were determined to be at least 95% pure by LC analysis.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(S)-2-(1-((9H-purin-6-yl)amino)propyl)-5-fluoro-3-(4-hydroxyphenyl)quinazolin-4(3H)-one (2). 2-Propanol (3
mL) was added to (S)-2-(1-aminopropyl)-5-fluoro-3-(4-hydroxyphenyl)quinazolin-4(3H)-one 2,2,2-trifluoroacetic
acid salt (32b) (0.25 mmol) and the solution was transferred to a sealed tube containing 6-chloro-9-(tetrahydro-2H-
pyran-2-yl)-9H-purine (33a) (60 mg, 0.25 mmol). Diisopropylethylamine (0.25 mL, 1.5 mmol) was added and the
reaction mixture was heated to 100 C for 16 hours. The reaction was cooled to room temperature and concentrated.
Dichloromethane (2.5 mL) and trifluoroacetic acid (0.39 mL, 5.0 mmol) were added. The solution was stirred at
room temperature for 3 hours and concentrated in vacuuo. A DMSO solution of the crude mixture was then purified
by HPLC eluting with 5-95% water/acetonitrile (0.1% v/v trifluoroacetic acid). The appropriate fractions were
pooled and lyophilized to afford (S)-2-(1-((9H-purin-6-yl)amino)propyl)-5-fluoro-3-(4-hydroxyphenyl)quinazolin-
1
4(3H)-one 2,2,2-trifluoroacetic acid salt (2) as a white solid (51 mg). H NMR (400 MHz, DMSO) δ 12.97 (s, 1H),
9.84 (s, 1H), 8.29 – 8.11 (m, 2H), 7.85 – 7.70 (m, 2H) 7.42 (d, J = 8.4 Hz, 2H), 7.35 – 7.40 (m, 2H), 6.95 – 6.80 (m,
2H), 4.74 (br s, 1H), 1.89 (br s, 2H), 0.76 (t, J = 7.2 Hz). ES/MS m/z = 432.3 (M+H)+.
ACS Paragon Plus Environment

Page 19 of 35
Journal of Medicinal Chemistry
1
2
3
4
General procedure A: (S)-4-amino-6-((1-(5-chloro-4-oxo-3-phenyl-3,4-dihydroquinazolin-2-
5
6
7
8
yl)ethyl)amino)pyrimidine-5-carbonitrile (3). 2-Propanol (3 mL) was added to (S)-2-(1-aminoethyl)-5-chloro-3-
phenylquinazolin-4(3H)-one 2,2,2-trifluoroacetic acid salt (32c) (0.25 mmol) and the solution was transferred to a
sealed tube containing 4-amino-6-chloropyrimidine-5-carbonitrile (33b) (39 mg, 0.25 mmol).
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Diisopropylethylamine (0.25 mL, 1.5 mmol) was added and the reaction mixture was heated to 80 C for 16 hours.
The reaction was cooled to room temperature and concentrated. A DMSO solution of the crude mixture was then
purified by HPLC eluting with 5-95% water/acetonitrile (0.1% v/v trifluoroacetic acid). The appropriate fractions
were pooled and lyophilized to afford (S)-4-amino-6-((1-(5-chloro-4-oxo-3-phenyl-3,4-dihydroquinazolin-2-
1
yl)ethyl)amino)pyrimidine-5-carbonitrile (3) as a white solid (53 mg). H NMR (400 MHz, DMSO) δ 7.85 (s, 1H),
7.80 − 7.40 (m, 8H), 7.25 (br, 2H), 4.75 (m, 1H), 1.33 (d, 3H). ES/MS m/z = 418.1 (M+H)+.
(S)-4-amino-6-((1-(5-chloro-3-(4-hydroxyphenyl)-4-oxo-3,4-dihydroquinazolin-2-yl)ethyl)amino)pyrimidine-
5-carbonitrile (4). Prepared according to general procedure A using 32d. 1H NMR (400 MHz, DMSO) δ 1H NMR
(400 MHz, DMSO-d6) δ 9.79 (s, 1H), 7.95 (s, 1H), 7.76 (t, J = 8.0 Hz, 1H), 7.61 (dd, J = 8.2, 1.2 Hz, 1H), 7.56 (dd,
J = 7.7, 1.2 Hz, 1H), 7.50 (d, J = 6.6 Hz, 1H), 7.28 (ddt, J = 7.0, 4.5, 2.6 Hz, 3H), 6.90 – 6.80 (m, 2H), 4.79 (p, J =
6.6 Hz, 1H), 1.35 (d, J = 6.7 Hz, 3H). ES/MS m/z = 434.1 (M+H)+.
(S)-4-amino-6-((1-(5-chloro-4-oxo-3-(1H-pyrazol-3-yl)-3,4-dihydroquinazolin-2-yl)ethyl)amino)pyrimidine-5-
carbonitrile (5). Prepared according to general procedure A using 32e. 1H NMR (400 MHz, DMSO) δ 13.14 (br s,
1H), 7.93 (s, 1H), 7.87 (d, J = 2 Hz, 1H), 7.75 (t, J = 8 Hz, 1H), 7.58 (dd, J = 8, 1 Hz), 7.54 (dd, J = 8, 1Hz, 1H),
7.43 (d, J = 7 Hz, 1H), 7.26 (s, 2H), 6.40 (d, J = 2 Hz, 1H), 4.82 (p, J = 7 Hz, 1H), 1.36 (d, J = 7 Hz, 3H). ES/MS
m/z = 408.1 (M+H)+.
(S)-4-amino-6-((1-(5-chloro-4-oxo-3-(1H-pyrazol-4-yl)-3,4-dihydroquinazolin-2-yl)ethyl)amino)pyrimidine-5-
carbonitrile (6). Prepared according to general procedure A using 32f. 1H NMR (400 MHz, DMSO) δ 8.04 (s, 1H),
7.84 (s, 2H), 7.75 (d, J = 8 Hz, 2H), 7.60 (dd, J = 8.4, 1.2 Hz, 1H), 7.57 (dd, J = 6.4, 1.2 Hz, 1H), 7.45 (bs, 2H), 4.94
(m, 1H), 1.38 (d, J = 6.8 Hz, 3H). ES/MS m/z = 408.1 (M+H)+.
General procedure B: (S)-2,4-diamino-6-((1-(5-chloro-4-oxo-3-(1H-pyrazol-3-yl)-3,4-dihydroquinazolin-2-
yl)ethyl)amino)pyrimidine-5-carbonitrile (7). 2-Propanol (3 mL) was added to (S)-2-(1-aminoethyl)-5-chloro-3-
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 35
1
2
3
4
5
6
7
8
9
(1H-pyrazol-3-yl)quinazolin-4(3H)-one 2,2,2-trifluoroacetic acid salt (32e) (0.25 mmol) and the resulting solution
was transferred to a microwave vial containing 2,4-diamino-6-chloropyrimidine-5-carbonitrile (33c) (42 mg, 0.25
mmol). Diisopropylethylamine (0.25 mL, 1.5 mmol) was added and the reaction mixture was heated to 130 C for
14 hours in the microwave. The reaction was cooled to room temperature and concentrated. A DMSO solution of the
crude mixture was then purified by HPLC eluting with 5-95% water/acetonitrile (0.1% v/v trifluoroacetic acid). The
appropriate fractions were pooled and lyophilized to afford (S)-2,4-diamino-6-((1-(5-chloro-4-oxo-3-(1H-pyrazol-3-
yl)-3,4-dihydroquinazolin-2-yl)ethyl)amino)pyrimidine-5-carbonitrile 2,2,2-trifluoroacetic acid salt (7) as a white
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
solid (60 mg). H NMR (400 MHz, DMSO) δ 13.20 (br s, 1H), 7.91 (d, J = 2.4 Hz, 1H), 7.84 – 7.77 (m, 1H), 7.65 –
7.59 (m, 2H), 6.55 – 6.52 (m, 1H), 4.85 – 4.75 (m, 1H), 1.41 (d, J = 6.7 Hz, 3H). ES/MS m/z = 423.1 (M+H)+.
(S)-2,4-diamino-6-(((5-chloro-4-oxo-3-(1H-pyrazol-3-yl)-3,4-dihydroquinazolin-2-
yl)(cyclopropyl)methyl)amino)pyrimidine-5-carbonitrile (8). Prepared according to general procedure B using
32g. 1H NMR (400 MHz, DMSO) δ 13.15 (br s, 1H), 7.90 – 7.74 (m, 2H), 7.69 – 7.61 (m, 2H), 6.38 (d, J = 2.3 Hz,
1H), 4.72 (t, J = 7.6 Hz, 1H), 1.47 – 1.37 (m, 1H), 0.53 – 0.36 (m, 3H), 0.24 – 0.14 (m, 1H). ES/MS m/z = 449.1
(M+H)+.
(S)-5-chloro-2-(1-((2,6-diaminopyrimidin-4-yl)amino)ethyl)-3-(1H-pyrazol-3-yl)quinazolin-4(3H)-one (9).
Prepared according to general procedure B using 32e and 6-chloropyrimidine-2,4-diamine (33d). 1H NMR (400
MHz, DMSO) δ 13.17 (s, 1H), 10.89 (s, 1H), 8.22 (d, J = 6.6 Hz, 1H), 7.89 (s, 1H), 7.78 – 7.70 (m, 1H), 7.60 – 7.52
(m, 2H), 7.11 (s, 1H), 7.02 (s, 1H), 6.66 (d, J = 2.4 Hz, 1H), 5.17 (s, 1H), 4.57 – 4.47 (m, 1H), 1.34 (d, J = 6.9 Hz,
3H). ES/MS m/z = 398.1 (M+H)+.
General procedure C: (S)-5-chloro-2-(1-((2,6-diamino-5-chloropyrimidin-4-yl)amino)ethyl)-3-(1H-pyrazol-3-
yl)quinazolin-4(3H)-one (10). 2-Propanol (3 mL) was added to (S)-2-(1-aminoethyl)-5-chloro-3-(1H-pyrazol-3-
yl)quinazolin-4(3H)-one 2,2,2-trifluoroacetic acid salt (32e) (0.25 mmol) and the solution was transferred to a
microwave vial containing 5-chloro-6-fluoropyrimidine-2,4-diamine (33e) (41 mg, 0.25 mmol).
Diisopropylethylamine (0.25 mL, 1.5 mmol) was added and the reaction mixture was heated to 130 C for 14 hours
in the microwave. The reaction was cooled to room temperature and concentrated. A DMSO solution of the crude
mixture was then purified by HPLC eluting with 5-95% water/acetonitrile (0.1% v/v trifluoroacetic acid). The
appropriate fractions were pooled and lyophilized to afford (S)-5-chloro-2-(1-((2,6-diamino-5-chloropyrimidin-4-
ACS Paragon Plus Environment

Page 21 of 35
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
yl)amino)ethyl)-3-(1H-pyrazol-3-yl)quinazolin-4(3H)-one (10) as a white solid (55 mg). H NMR (400 MHz,
DMSO) δ 13.22 (br s, 1H), 7.91 (d, J = 2.4 Hz, 1H), 7.81 (t, J = 8.0 Hz, 1H), 7.71 (d, J = 6.8 Hz, 1H), 7.66 (dd, J =
8.2, 1.2 Hz, 1H), 7.62 (dd, J = 7.8, 1.2 Hz, 1H), 7.52 (br s, 2H), 7.38 (br s, 2H), 6.54 (d, J = 2.4 Hz, 1H), 4.79 (p, J
= 6.7 Hz, 1H), 1.43 (d, J = 6.8 Hz, 3H). ES/MS 432.1 (M+H)+.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(S)-5-chloro-2-(cyclopropyl((2,6-diamino-5-chloropyrimidin-4-yl)amino)methyl)-3-(1H-pyrazol-3-
yl)quinazolin-4(3H)-one (11). Prepared according to general procedure C using 32g. 1H NMR (400 MHz, DMSO)
δ 13.17 (br s, 1H), 7.86 (d, J = 2.3 Hz, 1H), 7.82 (t, J = 8.0 Hz, 1H), 7.70 (dd, J = 8.2, 1.2 Hz, 1H), 7.63 (dd, J =
7.8, 1.2 Hz, 1H), 7.49 (br s, 3H), 7.35 (br s, 2H), 6.38 (d, J = 2.3 Hz, 1H), 4.71 (t, J = 7.7 Hz, 1H), 1.42 – 1.32 (m,
1H), 0.53 – 0.37 (m, 3H), 0.22 – 0.13 (m, 1H). ES/MS m/z = 458.1 (M+H)+.
(S)-2-amino-4-(((5-chloro-4-oxo-3-(1H-pyrazol-3-yl)-3,4-dihydroquinazolin-2-yl)(cyclopropyl)methyl)amino)-
6-methylpyrimidine-5-carbonitrile (12). 2-Propanol (3 mL) was added to (S)-2-(amino(cyclopropyl)methyl)-5-
chloro-3-(1H-pyrazol-3-yl)quinazolin-4(3H)-one 2,2,2-trifluoroacetic acid salt (32g) (0.25 mmol) and the resulting
solution was transferred to a microwave vial containing 2-amino-4-chloro-6-methylpyrimidine-5-carbonitrile (33f)
(41 mg, 0.25 mmol). Diisopropylethylamine (0.25 mL, 1.5 mmol) was added and the reaction mixture was heated to
90 C for 16 hours in the microwave. The reaction was cooled to room temperature and concentrated. A DMSO
solution of the crude mixture was then purified by HPLC eluting with 5-95% water/acetonitrile (0.1% v/v
trifluoroacetic acid). The appropriate fractions were pooled and lyophilized to afford (S)-2-amino-4-(((5-chloro-4-
oxo-3-(1H-pyrazol-3-yl)-3,4-dihydroquinazolin-2-yl)(cyclopropyl)methyl)amino)-6-methylpyrimidine-5-
1
carbonitrile 2,2,2-trifluoroacetic acid salt (12) as a white solid (44 mg). H NMR (400 MHz, DMSO) δ 13.11 (br s,
1H), 7.87 – 7.79 (m, 2H), 7.67 (dd, J = 8.2, 1.2 Hz, 1H), 7.62 (dd, J = 7.8, 1.2 Hz, 1H), 6.38 (d, J = 2.3 Hz, 1H),
4.73 (t, J = 7.6 Hz, 1H), 2.30 (s, 3H), 1.47 – 1.37 (m, 1H), 0.53 – 0.37 (m, 3H), 0.24 – 0.15 (m, 1H). ES/MS m/z =
448.1 (M+H)+.
(S)-2-amino-4-chloro-6-(((5-chloro-4-oxo-3-(1H-pyrazol-3-yl)-3,4-dihydroquinazolin-2-
yl)(cyclopropyl)methyl)amino)pyrimidine-5-carbonitrile (13). 2-Propanol (3 mL) was added to (S)-2-
(amino(cyclopropyl)methyl)-5-chloro-3-(1H-pyrazol-3-yl)quinazolin-4(3H)-one 2,2,2-trifluoroacetic acid salt (32g)
(0.25 mmol) and the solution was transferred to a sealed tube containing 2-amino-4,6-dichloropyrimidine-5-
carbonitrile (33g) (47 mg, 0.25 mmol). Diisopropylethylamine (0.25 mL, 1.5 mmol) was added and the reaction
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 35
1
2
3
4
5
6
7
8
9
mixture was heated to 130 C for 16 hours. The reaction was cooled to room temperature and concentrated. A
DMSO solution of the crude mixture was then purified by HPLC eluting with 5-95% water/acetonitrile (0.1% v/v
trifluoroacetic acid). The appropriate fractions were pooled and lyophilized to afford (S)-2-amino-4-chloro-6-(((5-
chloro-4-oxo-3-(1H-pyrazol-3-yl)-3,4-dihydroquinazolin-2-yl)(cyclopropyl)methyl)amino)pyrimidine-5-carbonitrile
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
2,2,2-trifluoroacetic acid salt (13) as a white solid (65 mg). H NMR (400 MHz, DMSO) δ 13.08 (br s, 1H), 7.85 –
7.78 (m, 2H), 7.67 (dd, J = 8.2, 1.2 Hz, 1H), 7.62 (dd, J = 7.8, 1.2 Hz, 1H), 7.57 (d, J = 7.4 Hz, 1H), 7.29 (br s, 1H),
6.35 (d, J = 2.4 Hz, 1H), 4.71 (t, J = 7.7 Hz, 1H), 1.48 – 1.39 (m, 1H), 0.54 – 0.38 (m, 3H), 0.22 – 0.13 (m, 1H).
ES/MS m/z = 468.1 (M+H)+.
General procedure D: (S)-2-amino-4-(((5-chloro-4-oxo-3-(1H-pyrazol-3-yl)-3,4-dihydroquinazolin-2-
yl)(cyclopropyl)methyl)amino)-6-(difluoromethyl)pyrimidine-5-carbonitrile (14). 2-Propanol (3 mL) was added
to (S)-2-(amino(cyclopropyl)methyl)-5-chloro-3-(1H-pyrazol-3-yl)quinazolin-4(3H)-one 2,2,2-trifluoroacetic acid
salt (32g) (0.25 mmol) and the solution was transferred to a sealed tube containing 2-amino-4-chloro-6-
(difluoromethyl)pyrimidine-5-carbonitrile (33h) (51 mg, 0.25 mmol). Diisopropylethylamine (0.25 mL, 1.5 mmol)
was added and the reaction mixture was heated to 80 C for 2 hours. The reaction was cooled to room temperature
and concentrated. A DMSO solution of the crude mixture was then purified by HPLC eluting with 5-95%
water/acetonitrile (0.1% v/v trifluoroacetic acid). The appropriate fractions were pooled and lyophilized to afford
(S)-2-amino-4-(((5-chloro-4-oxo-3-(1H-pyrazol-3-yl)-3,4-dihydroquinazolin-2-yl)(cyclopropyl)methyl)amino)-6-
1
(difluoromethyl)pyrimidine-5-carbonitrile 2,2,2-trifluoroacetic acid salt (14) as a white solid (59 mg). H NMR
(400 MHz, DMSO) δ 13.10 (br s, 1H), 7.82 – 7.75 (m, 2H), 7.63 (dd, J = 8.2, 1.2 Hz, 1H), 7.59 (dd, J = 7.8, 1.2 Hz,
1H), 7.54 (br s, 1H), 7.42 (d, J = 7.4 Hz, 1H), 7.22 (br s, 1H), 6.63 (t, J = 53.6 Hz, 1H), 6.34 (d, J = 2.4 Hz, 1H),
4.74 (t, J = 7.5 Hz, 1H), 1.43 – 1.32 (m, 1H), 0.48 – 0.34 (m, 3H), 0.19 – 0.11 (m, 1H). ES/MS m/z = 484.1 (M+H)+.
(S)-2-amino-4-((1-(5-chloro-4-oxo-3-(1H-pyrazol-3-yl)-3,4-dihydroquinazolin-2-yl)ethyl)amino)-6-
(trifluoromethyl)pyrimidine-5-carbonitrile (15). 2-Propanol (3 mL) was added to (S)-2-(1-aminoethyl)-5-chloro-
3-(1H-pyrazol-3-yl)quinazolin-4(3H)-one 2,2,2-trifluoroacetic acid salt (32e) (0.25 mmol) and the resulting solution
was transferred to a sealed tube containing 2-amino-4-chloro-6-(trifluoromethyl)pyrimidine-5-carbonitrile (33i) (56
mg, 0.25 mmol). Diisopropylethylamine (0.25 mL, 1.5 mmol) was added and the reaction mixture was heated to 80
C for 1 hour. The reaction was cooled to room temperature and concentrated. A DMSO solution of the crude
ACS Paragon Plus Environment

Page 23 of 35
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
mixture was then purified by HPLC eluting with 5%-95% water/acetonitrile (0.1%v/v trifluoroacetic acid). The
appropriate fractions were pooled and lyophilized to afford (S)-2-amino-4-((1-(5-chloro-4-oxo-3-(1H-pyrazol-3-yl)-
3,4-dihydroquinazolin-2-yl)ethyl)amino)-6-(trifluoromethyl)pyrimidine-5-carbonitrile 2,2,2-trifluoroacetic acid salt
1
(15) as a white solid (57 mg). H NMR (400 MHz, DMSO) δ 13.17 (s, 1H), 7.99 (d, J = 6.5 Hz, 1H), 7.87 (d, J = 2.4
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Hz, 1H), 7.84 – 7.78 (m, 2H), 7.66 – 7.59 (m, 2H), 7.48 (br s, 1H), 6.53 (d, J = 2.4 Hz, 1H), 4.86 (p, J = 6.7 Hz, 1H),
1.45 (d, J = 6.7 Hz, 3H). ES/MS m/z = 476.1 (M+H)+.
(S)-2-(1-((2,6-diamino-5-chloropyrimidin-4-yl)amino)ethyl)-5-fluoro-3-(1H-pyrazol-3-yl)quinazolin-4(3H)-one
(16a). Prepared according to general procedure C using 32h. 1H NMR (400 MHz, DMSO) δ 13.21 (br s, 1H), 7.90
(d, J = 2.3 Hz, 1H), 7.89 – 7.82 (m, 1H), 7.71 (d, J = 6.7 Hz, 1H), 7.56 – 7.48 (m, 2H), 7.43 – 7.28 (m, 4H), 6.53 (d,
J = 2.4 Hz, 1H), 4.79 (p, J = 6.7 Hz, 1H), 1.43 (d, J = 6.8 Hz, 3H). ES/MS 416.1 (M+H)+.
(S)-2-amino-4-(difluoromethyl)-6-((1-(5-fluoro-4-oxo-3-(1H-pyrazol-3-yl)-3,4-dihydroquinazolin-2-
yl)ethyl)amino)pyrimidine-5-carbonitrile (16b). Prepared according to general procedure D using 32h. 1H NMR
(400 MHz, DMSO) δ 13.18 (br s, 1H), 7.90 – 7.82 (m, 2H), 7.80 (d, J = 6.4 Hz, 1H), 7.58 (br s, 1H), 7.52 – 7.47 (m,
1H), 7.34 (ddd, J = 11.0, 8.2, 1.0 Hz, 1H), 7.23 (br s, 1H), 6.67 (t, J = 53.6 Hz, 1H), 6.55 (d, J = 2.4 Hz, 1H), 4.80 (p,
J = 6.6 Hz, 1H), 1.44 (d, J = 6.7 Hz, 3H). ES/MS 442.0 (M+H)+.
(S)-2-(1-((2,6-diamino-5-chloropyrimidin-4-yl)amino)ethyl)-6-fluoro-3-(1H-pyrazol-3-yl)quinazolin-4(3H)-one
(17a). Prepared according to general procedure C using 32i. 1H NMR (400 MHz, DMSO) δ 13.21 (s, 1H), 7.92 (dd,
J = 2.4, 1.7 Hz, 1H), 7.85 – 7.63 (m, 2H), 6.57 (t, J = 2.2 Hz, 1H), 6.25 (d, J = 7.4 Hz, 1H), 5.96 (s, 2H), 5.50 (s,
2H), 4.62 (p, J = 6.8 Hz, 1H), 1.34 (d, J = 6.7 Hz, 3H). ES/MS 416.1 (M+H)+.
(S)-2-amino-4-(difluoromethyl)-6-((1-(6-fluoro-4-oxo-3-(1H-pyrazol-3-yl)-3,4-dihydroquinazolin-2-
yl)ethyl)amino)pyrimidine-5-carbonitrile (17b). Prepared according to general procedure D using 32i. 1H NMR
(400 MHz, DMSO) δ 13.19 (br s, 1H), 7.87 (d, J = 2.4 Hz, 1H), 7.85 – 7.74 (m, 4H), 7.57 (br s, 1H), 7.23 (br s, 1H),
6.67 (t, J = 53.6 Hz, 1H), 6.55 (d, J = 2.4 Hz, 1H), 4.87 (p, J = 6.6 Hz, 1H), 1.45 (d, J = 6.7 Hz, 3H). ES/MS 442.1
(M+H)+.
(S)-2-(1-((2,6-diamino-5-chloropyrimidin-4-yl)amino)ethyl)-8-fluoro-3-(1H-pyrazol-3-yl)quinazolin-4(3H)-one
(18a). Prepared according to general procedure C using 32j. 1H NMR (400 MHz, DMSO) δ 13.24 (br s, 1H), 7.98 –
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 35
1
2
3
4
5
6
7
8
7.94 (m, 1H), 7.91 (d, J = 2.4 Hz, 1H), 7.78 (ddd, J = 10.5, 8.1, 1.4 Hz, 1H), 7.72 (d, J = 6.7 Hz, 1H), 7.58 (td, J =
8.1, 4.7 Hz, 1H), 7.48 (br s, 2H), 7.33 (br s, 2H), 6.54 (d, J = 2.3 Hz, 1H), 4.85 (p, J = 6.6 Hz, 1H), 1.45 (d, J = 6.7
Hz, 3H). ES/MS 416.1 (M+H)+.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(S)-2-amino-4-(difluoromethyl)-6-((1-(8-fluoro-4-oxo-3-(1H-pyrazol-3-yl)-3,4-dihydroquinazolin-2-
yl)ethyl)amino)pyrimidine-5-carbonitrile (18b). Prepared according to general procedure D using 32j. 1H NMR
(400 MHz, DMSO) δ 13.19 (br s, 1H), 7.97 – 7.93 (m, 1H), 7.87 (d, J = 2.4 Hz, 1H), 7.83 (d, J = 6.5 Hz, 1H), 7.77
(ddd, J = 10.5, 8.0, 1.4 Hz, 1H), 7.60 – 7.53 (m, 2H), 7.24 (br s, 1H), 6.67 (t, J = 53.5 Hz, 1H), 6.54 (d, J = 2.4 Hz,
1H), 4.88 (p, J = 6.7 Hz, 1H), 1.46 (d, J = 6.7 Hz, 3H). ES/MS 442.1 (M+H)+.
(S)-2-(1-((2,6-diamino-5-chloropyrimidin-4-yl)amino)ethyl)-6,8-difluoro-3-(1H-pyrazol-3-yl)quinazolin-4(3H)-
one (19a). Prepared according to general procedure C using 32k. 1H NMR (400 MHz, DMSO) δ 13.22 (br s, 1H),
7.95 – 7.85 (m, 1H), 7.87 (d, J = 2.3 Hz, 1H), 7.74 – 7.69 (m, 1H), 7.67 (ddd, J = 8.3, 3.0, 1.3 Hz, 1H), 7.50 (br s,
2H), 7.36 (br s, 2H), 6.49 (d, J = 2.4 Hz, 1H), 4.83 (p, J = 6.7 Hz, 1H), 1.41 (d, J = 6.7 Hz, 3H). ES/MS 434.1
(M+H)+.
(S)-2-amino-4-((1-(6,8-difluoro-4-oxo-3-(1H-pyrazol-3-yl)-3,4-dihydroquinazolin-2-yl)ethyl)amino)-6-
(difluoromethyl)pyrimidine-5-carbonitrile (19b). Prepared according to general procedure D using 32k. 1H NMR
(400 MHz, DMSO) δ 13.21 (br s, 1H), 7.92 (ddd, J = 10.3, 8.9, 2.9 Hz, 1H), 7.87 (d, J = 2.4 Hz, 1H), 7.82 (d, J = 6.5
Hz, 1H), 7.70 (ddd, J = 8.2, 2.9, 1.2 Hz, 1H), 7.57 (br s, 1H), 7.23 (br s, 1H), 6.67 (t, J = 53.5 Hz, 1H), 6.54 (d, J =
2.4 Hz, 1H), 4.88 (p, J = 6.7 Hz, 1H), 1.45 (d, J = 6.7 Hz, 3H). ES/MS 460.1 (M+H)+.
(S)-8-chloro-2-(1-((2,6-diamino-5-chloropyrimidin-4-yl)amino)ethyl)-6-fluoro-3-(1H-pyrazol-3-yl)quinazolin-
4(3H)-one (20a). Prepared according to general procedure C using 32l. 1H NMR (400 MHz, DMSO) δ 13.26 (s,
1H), 8.10 (dd, J = 8.5, 2.9 Hz, 1H), 7.94 (dd, J = 2.4, 1.6 Hz, 1H), 7.78 (dd, J = 8.2, 2.9 Hz, 1H), 6.53 (t, J = 2.2 Hz,
1H), 6.39 (d, J = 7.5 Hz, 1H), 5.97 (s, 2H), 5.55 (s, 2H), 4.71 (dd, J = 7.4, 6.6 Hz, 1H), 1.33 (d, J = 6.6 Hz, 3H).
ES/MS 450.1 (M+H)+.
(S)-2-amino-4-((1-(8-chloro-6-fluoro-4-oxo-3-(1H-pyrazol-3-yl)-3,4-dihydroquinazolin-2-yl)ethyl)amino)-6-
(difluoromethyl)pyrimidine-5-carbonitrile (20b). Prepared according to general procedure D using 32l. 1H NMR
(400 MHz, DMSO) δ 13.19 (s, 1H), 8.11 (dd, J = 8.5, 2.9 Hz, 1H), 7.86 (d, J = 2.4 Hz, 1H), 7.80 (dd, J = 8.2, 2.9
ACS Paragon Plus Environment

Page 25 of 35
Journal of Medicinal Chemistry
1
2
3
4
5
6
Hz, 1H), 7.72 (d, J = 6.6 Hz, 1H), 7.54 (br s, 1H), 7.30 (br s, 1H), 6.64 (t, J = 53.6 Hz, 1H), 6.46 (d, J = 2.4 Hz, 1H),
4.94 (p, J = 6.6 Hz, 1H), 1.42 (d, J = 6.6 Hz, 3H). ES/MS 476.1 (M+H)+.
7
8
9
(S)-8-chloro-2-(cyclopropyl((2,6-diamino-5-chloropyrimidin-4-yl)amino)methyl)-6-fluoro-3-(1H-pyrazol-3-
yl)quinazolin-4(3H)-one (21a). Prepared according to general procedure C using 32m. 1H NMR (400 MHz,
DMSO) δ 13.32 (d, J = 2.5 Hz, 1H), 8.14 (ddd, J = 8.6, 3.0, 1.7 Hz, 1H), 7.98 (t, J = 2.1 Hz, 1H), 7.82 (ddd, J = 8.2,
3.0, 1.7 Hz, 1H), 6.52 (q, J = 2.0 Hz, 1H), 6.24 – 5.89 (m, 3H), 5.61 (s, 2H), 4.92 (ddd, J = 7.9, 5.6, 1.8 Hz, 1H),
1.25 (qd, J = 7.3, 6.6, 3.8 Hz, 1H), 0.43 – 0.10 (m, 4H). ES/MS 476.1 (M+H)+.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(S)-2-amino-4-(((8-chloro-6-fluoro-4-oxo-3-(1H-pyrazol-3-yl)-3,4-dihydroquinazolin-2-
yl)(cyclopropyl)methyl)amino)-6-(difluoromethyl)pyrimidine-5-carbonitrile (21b). Prepared according to
general procedure D using 32m. 1H NMR (400 MHz, DMSO) δ 13.15 (s, 1H), 8.13 (dd, J = 8.5, 2.9 Hz, 1H), 7.84 –
7.78 (m, 2H), 7.55 (br s, 1H), 7.37 (d, J = 7.5 Hz, 1H), 7.29 (br s, 1H), 6.64 (t, J = 53.6 Hz, 1H), 6.35 (d, J = 2.4 Hz,
1H), 4.95 (t, J = 7.2 Hz, 1H), 1.48 – 1.34 (m, 1H), 0.48 – 0.33 (m, 3H), 0.19 – 0.10 (m, 1H). ES/MS 502.1 (M+H)+.
(S)-8-chloro-2-(1-((2,6-diamino-5-chloropyrimidin-4-yl)amino)propyl)-6-fluoro-3-(1H-pyrazol-3-
yl)quinazolin-4(3H)-one (22a). Prepared according to general procedure C using 32n. 1H NMR (400 MHz, DMSO)
δ 13.31 (d, J = 25.8 Hz, 1H), 8.18 (ddd, J = 12.4, 8.5, 2.9 Hz, 1H), 8.02 – 7.77 (m, 2H), 7.59 (d, J = 30.8 Hz, 2H),
7.41 (s, 2H), 6.54 (dd, J = 24.8, 2.4 Hz, 1H), 4.86 (td, J = 7.1, 4.6 Hz, 1H), 2.18 – 1.65 (m, 2H), 0.78 (dt, J = 14.4,
7.3 Hz, 3H). ES/MS 464.1 (M+H)+.
(S)-2-amino-4-((1-(8-chloro-6-fluoro-4-oxo-3-(1H-pyrazol-3-yl)-3,4-dihydroquinazolin-2-yl)propyl)amino)-6-
(difluoromethyl)pyrimidine-5-carbonitrile (22b). Prepared according to general procedure D using 32n. 1H NMR
(400 MHz, DMSO) δ 13.24 (br s, 1H), 8.14 (dd, J = 8.5, 2.9 Hz, 1H), 7.89 (d, J = 2.4 Hz, 1H), 7.83 (dd, J = 8.2, 2.9
Hz, 1H), 7.62 (d, J = 7.1 Hz, 1H), 7.58 (br s, 1H), 7.23 (br s, 1H), 6.68 (t, J = 53.5 Hz, 1H), 6.52 (d, J = 2.4 Hz, 1H),
4.94 – 4.88 (m, 1H), 2.12 – 1.99 (m, 1H), 1.78 (dq, J = 14.1, 7.1 Hz, 1H), 0.79 (t, J = 7.3 Hz, 3H). ES/MS 490.1
(M+H)+.
(S)-5-chloro-2-(1-((2,6-diamino-5-chloropyrimidin-4-yl)amino)ethyl)-3-(isoxazol-3-yl)quinazolin-4(3H)-one
(25). Prepared according to general procedure C using 32o. 1H NMR (400 MHz, DMSO) δ 9.10 (d, J = 1.7 Hz, 1H),
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 35
1
2
3
4
5
6
7.82 (t, J = 8.0 Hz, 1H), 7.73 – 7.56 (m, 2H), 7.55 – 6.98 (m, 5H), 6.88 (d, J = 1.8 Hz, 1H), 4.96 – 4.78 (m, 1H),
1.43 (d, J = 6.7 Hz, 3H). ES/MS 433.1 (M+H)+.
7
8
9
(S)-2,4-diamino-6-(((8-chloro-6-fluoro-4-oxo-3-(1H-pyrazol-3-yl)-3,4-dihydroquinazolin-2-
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
yl)(cyclopropyl)methyl)amino)pyrimidine-5-carbonitrile (26). Prepared according to general procedure B using
32l. 1H NMR (400 MHz, DMSO) δ 13.38 – 13.15 (m, 1H), 8.14 (dd, J = 8.5, 2.9 Hz, 1H), 7.94 (dd, J = 2.4, 1.6 Hz,
1H), 7.82 (dd, J = 8.1, 2.9 Hz, 1H), 6.65 (s, 2H), 6.52 – 6.42 (m, 2H), 6.36 (s, 2H), 4.96 (dd, J = 7.7, 6.2 Hz, 1H),
1.43 – 1.26 (m, 1H), 0.50 – 0.29 (m, 3H), 0.18 (tt, J = 8.5, 3.3 Hz, 1H). ES/MS m/z = 467.1 (M+H)+.
General procedure E: tert-butyl (S)-(1-(5-chloro-4-oxo-3-(1H-pyrazol-3-yl)-3,4-dihydroquinazolin-2-
yl)ethyl)carbamate (31e). Diphenyl phosphite (1.6 mL, 8.2 mmol) was added to a solution of 2-amino-6-
chlorobenzoic acid (27b) (400 mg, 2.3 mmol) and N-(tert-butoxycarbonyl)-L-alanine (28b) (568 mg, 3.0 mmol) in
pyridine (2 mL). The reaction mixture was stirred at 40 C for 2 hours. 3-Amino-1H-pyrazole (30c) (232 mg, 2.8
mmol) was then added to the mixture, which was then stirred at 50 C for 16 hours. The reaction mixture was cooled
to room temperature. This mixture was diluted with EtOAc (50 mL), washed with saturated aqueous NaHCO3 (1 x
50 mL), brine (50 mL), and dried over magnesium sulfate. The organic layer was filtered and concentrated in
vacuuo to afford material which was purified by column chromatography on SiO2 eluting with EtOAc in hexanes
(10-70%) to afford the title compound tert-butyl (S)-(1-(5-chloro-4-oxo-3-(1H-pyrazol-3-yl)-3,4-dihydroquinazolin-
2-yl)ethyl)carbamate (31e) (789 mg, 88%). ES/MS m/z = 390.1 (M+H)+.
Intermediates 31a-o made using general procedure E: tert-butyl (S)-(1-(5-fluoro-4-oxo-3-phenyl-3,4-
dihydroquinazolin-2-yl)propyl)carbamate (31a); tert-butyl (S)-(1-(3-(4-((tert-butyldimethylsilyl)oxy)phenyl)-5-
fluoro-4-oxo-3,4-dihydroquinazolin-2-yl)propyl)carbamate (31b); tert-butyl (S)-(1-(5-chloro-4-oxo-3-phenyl-3,4-
dihydroquinazolin-2-yl)ethyl)carbamate (31c); tert-butyl (S)-(1-(3-(4-((tert-butyldimethylsilyl)oxy)phenyl)-5-
chloro-4-oxo-3,4-dihydroquinazolin-2-yl)ethyl)carbamate (31d); tert-butyl (S)-(1-(5-chloro-4-oxo-3-(1H-pyrazol-3-
yl)-3,4-dihydroquinazolin-2-yl)ethyl)carbamate (31e); tert-butyl (S)-(1-(5-chloro-4-oxo-3-(1H-pyrazol-4-yl)-3,4-
dihydroquinazolin-2-yl)ethyl)carbamate (31f); tert-butyl (S)-((5-chloro-4-oxo-3-(1H-pyrazol-3-yl)-3,4-
dihydroquinazolin-2-yl)(cyclopropyl)methyl)carbamate (31g); tert-butyl (S)-(1-(5-fluoro-4-oxo-3-(1H-pyrazol-3-
yl)-3,4-dihydroquinazolin-2-yl)ethyl)carbamate (31h); tert-butyl (S)-(1-(6-fluoro-4-oxo-3-(1H-pyrazol-3-yl)-3,4-
dihydroquinazolin-2-yl)ethyl)carbamate (31i); tert-butyl (S)-(1-(8-fluoro-4-oxo-3-(1H-pyrazol-3-yl)-3,4-
ACS Paragon Plus Environment

Page 27 of 35
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
dihydroquinazolin-2-yl)ethyl)carbamate (31j); tert-butyl (S)-(1-(6,8-difluoro-4-oxo-3-(1H-pyrazol-3-yl)-3,4-
dihydroquinazolin-2-yl)ethyl)carbamate (31k); tert-butyl (S)-(1-(8-chloro-6-fluoro-4-oxo-3-(1H-pyrazol-3-yl)-3,4-
dihydroquinazolin-2-yl)ethyl)carbamate (31l); tert-butyl (S)-((8-chloro-6-fluoro-4-oxo-3-(1H-pyrazol-3-yl)-3,4-
dihydroquinazolin-2-yl)(cyclopropyl)methyl)carbamate (31m); tert-butyl (S)-(1-(8-chloro-6-fluoro-4-oxo-3-(1H-
pyrazol-3-yl)-3,4-dihydroquinazolin-2-yl)propyl)carbamate (31n); tert-butyl (S)-(1-(5-chloro-3-(isoxazol-3-yl)-4-
oxo-3,4-dihydroquinazolin-2-yl)ethyl)carbamate (31o).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
General procedure F: (S)-2-(1-aminoethyl)-5-chloro-3-(1H-pyrazol-3-yl)quinazolin-4(3H)-one (32e).
Trifluoroacetic acid (0.39 mL, 5.0 mmol) was added to a suspension of tert-butyl (S)-(1-(5-chloro-4-oxo-3-(1H-
pyrazol-3-yl)-3,4-dihydroquinazolin-2-yl)ethyl)carbamate (31e) (100 mg, 0.25 mmol) in dichloromethane (2.5 mL).
The solution was stirred at room temperature for 3 hours and concentrated in vacuuo to afford (S)-2-(1-aminoethyl)-
5-chloro-3-(1H-pyrazol-3-yl)quinazolin-4(3H)-one 2,2,2-trifluoroacetic acid salt (32e) as a gum which was used
without any further purification. ES/MS m/z = 290.1 (M+H)+.
Intermediates 32a-o made using general procedure F: (S)-2-(1-aminopropyl)-5-fluoro-3-phenylquinazolin-
4(3H)-one (32a); (S)-2-(1-aminopropyl)-5-fluoro-3-(4-hydroxyphenyl)quinazolin-4(3H)-one (32b); (S)-2-(1-
aminoethyl)-5-chloro-3-phenylquinazolin-4(3H)-one (32c); (S)-2-(1-aminoethyl)-5-chloro-3-(4-
hydroxyphenyl)quinazolin-4(3H)-one (32d); (S)-2-(1-aminoethyl)-5-chloro-3-(1H-pyrazol-3-yl)quinazolin-4(3H)-
one (32e); (S)-2-(1-aminoethyl)-5-chloro-3-(1H-pyrazol-4-yl)quinazolin-4(3H)-one (32f); (S)-2-
(amino(cyclopropyl)methyl)-5-chloro-3-(1H-pyrazol-3-yl)quinazolin-4(3H)-one (32g); (S)-2-(1-aminoethyl)-5-
fluoro-3-(1H-pyrazol-3-yl)quinazolin-4(3H)-one (32h); (S)-2-(1-aminoethyl)-6-fluoro-3-(1H-pyrazol-3-
yl)quinazolin-4(3H)-one (32i); (S)-2-(1-aminoethyl)-8-fluoro-3-(1H-pyrazol-3-yl)quinazolin-4(3H)-one (32j); (S)-2-
(1-aminoethyl)-6,8-difluoro-3-(1H-pyrazol-3-yl)quinazolin-4(3H)-one (32k); (S)-2-(1-aminoethyl)-8-chloro-6-
fluoro-3-(1H-pyrazol-3-yl)quinazolin-4(3H)-one (32l); (S)-2-(amino(cyclopropyl)methyl)-8-chloro-6-fluoro-3-(1H-
pyrazol-3-yl)quinazolin-4(3H)-one (32m); (S)-2-(1-aminopropyl)-8-chloro-6-fluoro-3-(1H-pyrazol-3-yl)quinazolin-
4(3H)-one (32n); (S)-2-(1-aminoethyl)-5-chloro-3-(isoxazol-3-yl)quinazolin-4(3H)-one (32o).
2,4-diamino-6-chloropyrimidine-5-carbonitrile (33c). Ammonium hydroxide (20 mL) was added to a solution of
2,4,6-trichloropyrimidine-5-carbonitrile (5.0 g, 24 mmol) in dioxane (20 mL) at room temperature. The solution was
warmed to 50 C and stirred for 3 hours. The reaction mixture was cooled to 10 C and water (50 mL) was added.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 35
1
2
3
4
The resulting solid was filtered, washed with water and dried under high vacuum to afford the title compound as a
white solid (4.5 g). 13C NMR (100 MHz, DMSO)  164.8, 162.6, 161.9, 115.8, 77.6. ES/MS m/z = 169.9 (M+H)+.
5
6
7
8
9
5-chloro-6-fluoropyrimidine-2,4-diamine (33e). To 2,4,6-trifluoropyrimidine (10 g, 75 mmol) in acetonitrile (100
mL) cooled to 0 C was added conc NH4OH (50 mL) in three portions. The reaction mixture was stirred at 40 °C
for 12 hours, after which time the reaction mixture was cooled to room temperature and concentrated in vacuuo. The
resulting solid was dissolved in MeOH/EtOH (250 mL, 1:1) and cooled to 0 oC. N-Chlorosuccinimide (13 g, 97
mmol) was added portion wise to the solution. The reaction mixture was stirred at 50 C for 12 hours. After which
time the cooled reaction mixture was concentrated to approximately 150 mL in vacuuo. The resultant suspension
was cooled to -10 °C and the solid formed was collected by filtration and washed with water (100 mL), followed by
0.1M NaOH (100 mL) and dried under high vacuum to afford the title compound as a white solid (8.7 g). ES/MS
m/z = 163.1 (M+H)+.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ASSOCIATED CONTENT
Supporting Information. Details of all protocols for in vitro and in vivo experiments, kinase selectivity
and molecular formula strings. This material is available free of charge via the Internet at
http://pubs.acs.org.
Accession Codes. The PI3Kβ homology model used was built from the crystal structure of (S)-2,4-
diamino-6-((5-chloro-8-fluoro-4-oxo-3-(pyridin-3-yl)-3,4-dihydroquinazolin-2-
yl)(cyclopropyl)methylamino)-pyrimidine-5-carbonitrile (GS-9901) bound to PI3K (pdb: 5T8I;
coordinates available in Supporting Information).12
AUTHOR INFORMATION
Corresponding Author. *E-mail: stephane.perreault@gilead.com. Phone: (206) 832-2029. Address:
Gilead Sciences, Inc., 199 E Blaine Street, Seattle, Washington 98102, United States.
Present Address.†J.E.: Acerta Pharma, 15400 SE 30th Place, Suite 206, Bellevue, WA 98007, U.S.
ACS Paragon Plus Environment

Page 29 of 35
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
Notes. The authors declare the following competing financial interest(s): The authors are employees of
Gilead Sciences except for J.E who was employed at Gilead Sciences during this research. All authors are
shareholders of Gilead Sciences.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ABREVIATIONS USED
PI3K, phosphoinositide 3-kinase; PTEN, phosphatase and tensin homolog; mTOR, mechanistic target of
rapamycin; AKT, protein kinase B; PIP2, phosphatidylinositol-4,5-bisphosphate; PIP3,
phosphatidylinositol-3,4,5-trisphosphate; pAKT, phosphorylated protein kinase B; tPSA, total polar
surface area; CLpr, predicted clearance;
(1) Fruman, D.A.; Rommel, C. PI3K and cancer: lessons, challenges and opportunities. Nat. Rev. Drug Discov.
2014, 13, 140-156.
(2) Sadeghi, N.; Gerber, D. E. Targeting the PI3K pathway for cancer therapy. Future Med. Chem. 2012, 4, 1153-
1169.
(3) Vanhaesebroeck, B.; Stephens, L.; Hawkins, P. PI3K signaling: the path to discovery and understanding. Nat.
Rev. Mol. Cell Biol. 2012, 13, 195-203.
(4) Liu, P.; Cheng, H.; Roberts, T. M.; Zhao, J. J. Targeting the phosphoinositide 3-kinase pathway in cancer. Nat.
Rev. Drug Discov. 2009, 8, 627−644.
(5) Yuan T. L.; Cantley L. C. PI3K pathway alterations in cancer: variations on a theme. Oncogene 2008, 27, 5497-
5510.
(6) Zhao, L.; Vogt, P. K. Class 1 PI3K in oncogenic cellular transformation. Oncogene 2008, 27, 5486−5496.
(7) Amzel, L. M.; Huang, C.-H.; Mandelker, D.; Lengauer, C.; Gabelli, S. B.; Vogelstein, B. Structural comparisons
of class I phosphoinositide 3-kinase. Nat. Rev. Cancer 2008, 8, 665-669.
(8) Courtney, K. D.; Corcoran, R. B.; Engelman, J. A. The PI3K pathway as drug target in human cancer. J. Clin.
Oncol. 2010, 28, 1075-1083.
(9) Rodon, J.; Dienstmann, R.; Serra, V.; Tabernero, J. Development of PI3K inhibitors: lessons learned from early
clinical trials. Nat. Rev. Clin. Oncol. 2013, 10, 143-153.
ACS Paragon Plus Environment