2,4,6-Triaminopyrimidine as a Novel Hinge Binder in a Series of PI3Kδ Selective Inhibitors

Article abstract of DOI:10.1021/acs.jmedchem.6b00213

Inhibition of phosphoinositide 3-kinase δ (PI3Kδ) is an appealing target for several hematological malignancies and inflammatory diseases. Herein, we describe the discovery and optimization of a series of propeller shaped PI3Kδ inhibitors comprising a novel triaminopyrimidine hinge binder. Combinations of electronic and structural strategies were employed to mitigate aldehyde oxidase mediated metabolism. This medicinal chemistry effort culminated in the identification of 52, a potent and highly selective inhibitor of PI3Kδ that demonstrates efficacy in a rat model of arthritis.

Full text of DOI:10.1021/acs.jmedchem.6b00213

Subscriber access provided by - Access paid by the | UCSB Libraries
Article
2,4,6-Triaminopyrimidine as a Novel Hinge
Binder in a Series of PI3K# Selective Inhibitors
Leena Patel, Jayaraman Chandrasekhar, Jerry B Evarts, Aaron Haran, Carmen Ip, Joshua
Kaplan, Musong Kim, David J Koditek, latesh lad, Eve-Irene Lepist, Mary E. McGrath, Nikolai
Novikov, Stéphane Perreault, Kamal D Puri, John R. Somoza, Bart H Steiner, Kirk Stevens,
Joseph Therrien, Jennifer Treiberg, Armando Garcia Villaseñor, Arthur Yeung, and Gary Phillips
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b00213 • Publication Date (Web): 16 Mar 2016
Downloaded from http://pubs.acs.org on March 20, 2016
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 47
Journal of Medicinal Chemistry
1
2
3
4
2,4,6-Triaminopyrimidine as a Novel Hinge Binder in a Series of PI3K Selective Inhibitors
5
6
7
8
9
Leena Patel,* Jayaraman Chandrasekhar, Jerry Evarts,† Aaron C. Haran, Carmen Ip, Joshua A. Kaplan,
Musong Kim, David Koditek, Latesh Lad, Eve-Irene Lepist, Mary E. McGrath, Nikolai Novikov,
Stephane Perreault, Kamal D. Puri,‡ John R. Somoza, Bart H. Steiner, Kirk L. Stevens, Joseph Therrien,
Jennifer Treiberg, Armando G. Villaseñor, Arthur Yeung and Gary Phillips
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
Gilead Sciences, Inc. 199 E Blaine St, Seattle, Washington 98102, United States
Gilead Sciences, Inc. 333 Lakeside Drive, Foster City, California 94404, United States
ABSTRACT
Inhibition of phosphoinositide-3-kinase delta (PI3K) is an appealing target for several hematological
malignancies and inflammatory diseases. Herein, we describe the discovery and optimization of a series
of propeller shaped PI3K inhibitors comprising of a novel triaminopyrimidine hinge binder.
Combinations of electronic and structural strategies were employed to mitigate aldehyde oxidase
mediated metabolism. This medicinal chemistry effort culminated in the identification of 52, a potent and
highly selective inhibitor of PI3Kthat demonstrates efficacy in a rat model of arthritis.
INTRODUCTION
Phosphoinositide-3-kinase delta (PI3K) is critical to the etiology of a number of diseases involving B-
cell function.1 Activation at the cell surface results in PI3K catalyzing the intracellular phosphorylation
of phosphatidylinositol 4,5-bisphosphate (PIP2) to the second messenger phosphatidylinositol 3,4,5-
trisphosphate (PIP3). PIP3 recruits cytosolic signaling enzymes such as Akt that contain pleckstrin
homology (PH) domains to the plasma membrane, triggering cell survival, growth and metabolism.2
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 47
1
2
3
4
PI3K is predominantly expressed in hematopoietic cells,3 and inhibition of PI3K has been shown to be
5
6
beneficial for the treatment of hematological malignancies where the PI3K/Akt pathway is hyperactive.4
7
8
9
Aberrant signaling in B-cells is also found in a number of inflammatory diseases. PI3K is highly
expressed in rheumatoid arthritis (RA) synovium,5 and inhibition of PI3K can reduce both B-cell and T-
cell production of inflammatory cytokines that may ameliorate the production of auto-antibodies. In
models of RA, PI3K inhibitors have demonstrated efficacy with respect to inflammation,6 and reduced
bone and cartilage erosion. In allergic asthma, PI3Kδ has been shown to be overactive in mast cells,7
neutrophils,8 T-cells,9 eosinophils, and B-cells.
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
Zydelig, (idelalisib), previously called GS-1101 or CAL-101 (1) (Figure 1a), is a first-in-class, oral,
potent and selective inhibitor of PI3KZydelig is dosed twice-daily and is approved for use in patients
with relapsed chronic lymphocytic leukemia (CLL), follicular lymphoma (FL) and small lymphocytic
lymphoma (SLL)10
As revealed by the crystal structure (Figure 1b), idelalisib (1) binds to the ATP site of the p110 catalytic
subunit of PI3K with the purine forming two hydrogen bonds with hinge residues Val828 and Glu826.10
The fluoro-quinazolinone ring assumes a perpendicular conformation to the hinge binder, and is
embedded in a specificity pocket between Trp760 and Met752 that is closed in the apo structure of the
enzyme.11 The phenyl ring lies perpendicular to the quinazolinone and protrudes out of the ATP-binding
pocket into an area known as hydrophobic region II (H2). The improved PI3K selectivity of propeller
shaped compounds such as idelalisib is presumably due to a lower energy requirement by PI3K for the
formation of the specificity pocket relative to the PI3K, PI3K, and PI3K isoforms.11b This previously-
described binding mode11a is also adopted by several compounds that have been evaluated in clinical
studies (Figure 2). Duvelisib, a delta-weighted  inhibitor, has demonstrated efficacy in B-cell and T-
cell lymphomas.12 AMG-31913 and TGR-1202 also inhibit the PI3K isoform and have been tested in
phase I clinical trials in CLL and non-Hodgkin lymphoma (NHL) patients.14
ACS Paragon Plus Environment

Page 3 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
In this manuscript, we describe our efforts toward identifying potent, highly selective and metabolically
stable PI3K inhibitors, suitable for the treatment of hematological malignancies and/or inflammatory
diseases at a low, once-daily oral dose.
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
(a)
(b)
Figure 1. (a) Structure of idelalisib10 (1) (b) 2.4 Å X-ray crystal structure of 1 bound in the PI3K ATP-binding site (PDB
code 4XE0). Idelalisib (1) binds in a propeller shape with the quinazolinone ring occupying an induced specificity pocket
formed by Trp760 and Met752. The N3 of the purine hinge binder interacts with the backbone NH of Val828 while the N9
forms short contacts with the backbone carbonyl of Glu826. Blue dashed lines show hydrogen bond contacts between the
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 47
1
2
3
4
5
inhibitor and protein. The Asn836 and Asp832 residues differ among the PI3K isoforms. For clarity, all other residues
have been removed.
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. PI3K delta inhibitors in clinical trials
CHEMISTRY
General synthetic route to novel quinazolinone compounds
Scheme 1a
a Reagents and Conditions: (i) Pyridine, HOP(OPh)2, 50 oC, 2 h, 100%; (ii) 50 oC 3-12 h, 50-100%; (iii) CF3COOH,
CH2Cl2, RT, 0.5 -4 h, 80-100%; (iv) i-Pr2NEt, IPA, 120 oC W or thermal, 8-14 h, 30-95%.
ACS Paragon Plus Environment

Page 5 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Analogs were synthesized following the general synthetic route shown in Scheme 1.15 The quinazolinone
core 6 was constructed via a one-pot, 2-step dehydrative cyclization. Oxazinone intermediate 4 was
prepared by treatment of anthranilic acid 2, with protected amino acid 3 in the presence of diphenyl
phosphite and pyridine. Subsequent addition of aniline 5 to the reaction mixture resulted in the formation
of the quinazolinone core 6. Deprotection of 6 with TFA afforded the primary amine intermediate 7 in
high yield. Incorporation of the hinge binder motif was attained by nucleophilic displacement of
substituted halo-heterocycles 8 in the presence of Hünigs base at high temperatures, either thermally or
under microwave irradiation.
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
Preparation of acetylene-containing compounds
Scheme 2a
a Reagents and Conditions: (i) Pd(PPh3)2Cl2, CuI, Et3N, Toluene, 80 oC, 3 h, 65-80%; (ii) TBAF, THF; (iii) Pd(PPh3)2Cl2,
CuI, Et3N, DMF, 90 oC, 3 h, 60-80%.
An alternative series of pyrimidines targeting the affinity pocket was synthesized according to the
sequence described in Scheme 2. Ethynyltrimethylsilane was functionalized via a Sonagashira coupling
with the corresponding aryl or heteroaryl bromide 10. Subsequent deprotection of the silyl protected
alkynes afforded compounds of the corresponding terminal alkyne 11. The synthesis was concluded
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 47
1
2
3
4
5
6
through a second Sonagashira coupling of 11 with intermediate 12 (synthesized according to the
procedure described in Scheme 1), to afford final compounds of the general structure 13 in good yields.
7
8
9
RESULTS AND DISCUSSION
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
To achieve our goal we set ourselves a target profile of selectivity greater than 150-fold for PI3K over
PI3K, PI3K, and PI3K. This would allow EC90 coverage of PI3K at trough concentrations without
compromising isoform selectivity. Additionally, a low in vitro human predicted clearance of 0.2 L/h/kg or
less was desired for a once-daily dosing regimen. Compound 14 served as an attractive starting point due
to its PI3Kbiochemicalpotency (PI3K IC50 = 1 nM). In an effort to optimize metabolic stability we
first focused on replacing the purine ring which is the primary site of metabolism for 1.16 We designed
compound 15 to retain the N3 pyrimidine nitrogen and the 4-amino moiety, both of which were believed
to be essential for potency. Unfortunately, this structural modification led to a 100-fold loss in PI3K
potency relative to compound 14. The loss in potency may be due to removal of the N7 of the purine ring,
which participates in a water-mediated hydrogen bonding network between the quinazolinone ring and
the protein.10 Removal of the 4-amino group (16) resulted in an additional loss of potency, illustrating the
essential nature of each hinge binding component. Incorporation of a 5-cyano group, exemplified by
compounds 17 and 18, restored PI3K potency with excellent selectivity over the other PI3K isoforms.
We synthesized the pyridine analog 19 which exhibited a 75-fold loss in potency, demonstrating the
importance of the N1 nitrogen. Finally, the regioisomer of 18, compound 20, resulted in lower potency
and diminished selectivity over PI3K. Due to its improved potency and selectivity profile, 18 was
advanced into drug metabolism and pharmacokinetic (DMPK) evaluation.
ACS Paragon Plus Environment

Page 7 of 47
Journal of Medicinal Chemistry
1
2
3
4
Table 1. Optimization of the Purine 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
a
Compound
Hinge
Isoform Selectivitya, b
PI3KIC50
nM



14
1
1200
290
55
15
16
17
99
1700
8
>100
>6
>100
>6
>100
>6
1200
460
370
18
0.4
4600
200
680
19
20
30
>340
530
>340
170
>340
15
2.5
aAverage of ≥ 1 determination. Run in quadruplicate. bDetermined from isoform specific PI3K IC50 values.
The predicted clearance of compound 18 in human liver microsomes (HLM) and rat liver microsomes
(RLM) were determined to be 0.1 L/h/kg and 0.2 L/h/kg, respectively. Encouraged by the low predicted
clearances we progressed 18 to in vivo pharmacokinetic evaluation (Table 2). When administered in rats
at 0.44 mg/kg intravenously, 18 demonstrated an intermediate in vivo systemic clearance of 1.1 L/h/kg
indicating a 5-fold difference between the observed in vivo clearance and the predicted in vitro
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 47
1
2
3
4
5
6
7
8
9
microsomal clearance. This discrepancy prompted us to assess susceptibility of the electron deficient 5-
cyano-pyrimidine ring of 18 to aldehyde oxidase (AO) metabolism. Aldehyde oxidase is a molybdenum-
containing cytoplasmic enzyme that catalyzes the oxidation of nitrogen-containing heterocycles.17
Microsomal assays utilize membrane fractions and do not capture clearance by AO.18 For that reason, the
stability of 18 was tested in human hepatocytes (hHeps) with and without raloxifene, a known inhibitor of
AO. A much higher predicted clearance of 0.88 L/h/kg in hHeps was determined for 18 relative to the 0.1
L/h/kg measured in the HLM assay. In the presence of raloxifene there was very little turnover of 18 in
the hHeps and the predicted clearance was similar to that observed in HLM (hHeps + raloxifene = 0.09
L/h/kg). Metabolite identification studies revealed the presence of an M+16 metabolite 21 (PI3K IC50 =
1500 nM), which was verified by independent synthesis (Figure 3). The identity of the product and
inhibition of metabolism in the presence of raloxifene are consistent with the involvement of AO in the
metabolism of 18.
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 2. DMPK Profile Compound 18
Compound
Species
Clearance (L/h/kg)
Vss (L/kg) b, c, #
2.6 ± 0.97
t1/2 (h) b, #
2.6 ± 1.7
in vitroa
in vivo b, #
1.1 ± 0.57
18
Rat
0.2
0.1
Human
b
aFrom hepatic microsomes. Intravenous dose of 0.44 mg/kg. cVolume of distribution at steady state. 18 was formulated in 5%
ethanol, 75% PEG 400, and 20% citrate buffer (pH 3.0) at 0.09 mg/mL and dosed in Sprague-Dawley rats. # Number of animals
per dosing route = 3.
ACS Paragon Plus Environment

Page 9 of 47
Journal of Medicinal Chemistry
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 3. Aldehyde Oxidase mediated metabolism of 18 to its M+16 metabolite 21.
Despite the high clearance of 18 in hHeps, the favorable potency and selectivity of this compound
prompted us to investigate alternative substitutions at the C5-position on the pyrimidine hinge binder. We
hoped to alleviate AO metabolism by modulating the electronics of the pyrimidine ring (Table 3).
Hydrolysis of the nitrile of 18 yielded the primary amide 22, which lost potency, but showed markedly
improved hepatic stability compared to 18. However, the stability of 22 improved further upon incubation
with raloxifene, indicating the involvement of AO. As anticipated, the trifluoromethyl pyrimidine 23 was
also a substrate for AO metabolism. We were hopeful that compound 24 would have diminished AO
metabolism due to the presence of a methyl group on the pyrimidine ring. Unfortunately, this
modification did not impact susceptibility of the hinge binder towards AO. In addition, 24 was much less
potent than 18, presumably due to weaker interactions of the pyrimidine with the hinge residues. Other
electron donating groups such as the methoxy substituent of 25 also resulted in a loss of potency.
Terminal acetylene 26 was synthesized to mimic the nitrile of compound 18 and exhibited excellent
PI3K potency and selectivity. Molecular modeling suggested that elaboration of the terminal acetylene
of 26 could allow exploration of the affinity pocket.11a The affinity pocket is accessed via a narrow
channel proximal to the hinge region and is lined with several conserved polar residues. Additional
docking studies on a number of analogs suggested that aryl and heteroaryl substituents on the acetylene of
26 may be able to interact with these residues through hydrogen bonding and/or cation interactions.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 47
1
2
3
4
Table 3. Effect of Modifying the Electronics of the Pyrimidine 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
a
Compound
Hinge
Isoform Selectivitya, b
hHepsc
L/h/kg
hHeps + raloxifenec
L/h/kg
PI3KIC50
nM



18
22
23
0.4
11
4600
200
680
310
560
0.88
0.27
0.59
0.09
0.08
0.12
860
1200
700
310
340
1.8
24
25
26
14
39
380
700
>251
1900
0.73
ND
0.25
ND
ND
>251
9700
>251
490
0.1
0.66
aAverage of ≥ 1 determination. Run in quadruplicate. bDetermined from isoform specific PI3K IC50 values. cND = Value not
determined.
To probe the affinity pocket we prepared analogs 27-32 (Table 4). Phenyl analog 27 exhibited good
inhibition of PI3Kbut diminished PI3K selectivity. Compounds 28 and 29, although potent, lost
selectivity over PI3K and PI3K, respectively. The pyridazine analog 30 suffered considerable potency
loss, possibly due to a repulsion between the heterocycle and nearby residues in the affinity pocketAn
overall improvement in potency and isoform selectivity was achieved with compounds 31 and 32. To
ACS Paragon Plus Environment

Page 11 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
better understand the potency of the affinity compounds, we examined an X-ray co-crystal of compound
32 bound to PI3K (Figure 4). The crystal structure illustrates the formation of two hydrogen bonds
between the aminopyrimidine and the hinge residues, Val828 and Glu826. In addition, the pyrazole of 32
forms short contacts with Lys779, Asp911 and Asp787 in the affinity pocket. These supplementary
hydrogen bonds help explain the gain in enzyme potency achieved with the acetylene-containing
compounds. Hopeful that the significant structural modifications relative to 18 would perturb AO
recognition of the hinge binder17d, 19 we quantified the metabolic fate for 31 in hHeps (+/– raloxifene).
Unfortunately, compound 31 exhibited a 4-fold reduction in predicted clearance when incubated with
raloxifene (0.13 L/h/kg); metabolite identification studies of 31 revealed an M+16 metabolite M1,
implicating AO metabolism (Figure 5). Despite marked improvement in isoform selectivity of 31 and 32,
the involvement of AO in the metabolism of these analogs deterred us from further exploration of
acetylene-containing compounds.
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. SAR Results for Acetylene-Containing Compounds
Isoform Selectivitya, b
a
PI3KIC50



Compound
27
R7
nM
0.3
1600
120
320
28
29
0.5
0.4
620
240
52
100
512
3200
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 47
1
2
3
4
5
6
7
8
9
30
31
32
240
1.2
0.3
>42
2600
1500
30
>42
470
230
330
260
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
aAverage of ≥ 1 determination. Run in quadruplicate. bDetermined from isoform specific PI3K IC50 values.
Figure 4. 2.85 Å X-ray crystal structure of 32 bound in the PI3K ATP-binding site (PDB accession code 5I6U). In
addition to the classic hinge binding interactions, the pyrazole ring forms hydrogen bonds to Lys779, Asp911 and Asp787
in the affinity pocket.
Figure 5. Aldehyde Oxidase mediated metabolism of 31 to its metabolite M1
ACS Paragon Plus Environment

Page 13 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Achieving metabolic stability was paramount in our program; electronic and structural modifications of
the hinge binder failed to alleviate AO metabolism. Therefore, we decided to focus on an alternate
strategy, directed at blocking oxidation of the C2-position on the pyrimidine ring. Deuteration at C2 (33)
modestly improved predicted clearance in hHeps compared to 18 (Table 5). Subsequent incubation of 33
with raloxifene revealed a 5-fold reduction in hepatic clearance implicating the involvement of AO.
Substitution with a primary amine (34) afforded high potency and low clearance that was insensitive to
the presence of raloxifene. The low predicted hepatic clearance of 34 was attractive but was accompanied
by an increase PI3K and PI3K potency. Other blocking groups such as chlorine (35) resulted in a loss
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
in potency, a further loss of PI3Kselectivity, and a higher predicted clearance relative to 34. The poor
metabolic stability is presumably due to an increase in lipophilicity (calculated LogP = 3.7 vs 3.0 (34)).
Similarly, other small substituent changes around the pyrimidine ring (36-38) furnished compounds that
exhibited high predicted clearances. The 2-amino substituent of 34 demonstrated a unique ability to
successfully block AO metabolism at the C2-position while retaining PI3K potency. The X-ray co-
crystal of 34 bound to PI3K reveals the triaminopyrimidine as a three-point hinge binding motif. In
addition to a hydrogen bond between the N3 and the backbone NH of Val828, a second hydrogen bond is
formed from the N4 to the backbone carbonyl of Glu826. The 2-amino substituent of the
triaminopyrimidine supplements the typical N3 and N4 hinge binding features with a third hydrogen bond
to the carbonyl of Val828 (Figure 6).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 47
1
2
3
4
Table 5. SAR Results for Fully Substituted Pyrimidines
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
a
Compound
Hinge
Isoform Selectivitya, b
hHeps
L/h/kg
hHeps + raloxifenec
PI3KIC50
L/h/kg
nM



18
33
34
35
0.4
1.2
0.1
4.3
4600
200
70
680
0.88
0.50
0.21
0.59
0.09
1200
2200
720
420
60
0.10
0.10
ND
80
180
37
36
37
38
0.9
1.0
2700
530
370
100
100
200
25
0.65
0.82
1.05
ND
ND
ND
0.13
1900
40
aAverage of ≥ 1 determination. Run in quadruplicate. bDetermined from isoform specific PI3K IC50 values. cND = Value not
determined.
ACS Paragon Plus Environment

Page 15 of 47
Journal of Medicinal Chemistry
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. 2.4 Å X-ray crystal structure of 34 in PI3K illustrates the triaminopyrimidine establishing hydrogen bonds to
the hinge residues, Val828 and Glu826. A third hydrogen bond is accomplished between the N2 component of the
triaminopyrimidine hinge binder and the carbonyl of Val828 (PDB accession code 5I4U).
Inspired by the potency and predicted clearance of 34, we focused our efforts upon improving isoform
selectivity. The H2 region of the p110 subunit of PI3K consists of two residues lining the edge of the
pocket that is filled by the phenyl ring. These residues differ among the PI3K isoforms; sequence
alignments reveal Asp832/Asn836 in PI3Kδ correspond to His/Gln in PI3Kα, Glu/Asp in PI3Kβ, and
Thr/Lys in PI3Kγ. As a result, the electrostatic and steric environments are significantly different across
the isoforms. We hypothesized that small substituents on the phenyl ring could influence isoform
selectivity. The addition of a fluoro substituent on the para position (39) resulted in significant loss of
isoform selectivity (Table 6). On the other hand, substitutions on the meta position (40-45) displayed
similar selectivity profiles to that of 34 with respect to PI3K. We thought that we may be able to further
improve the isoform selectivity of compound 42 by the addition of a second group to the 5-position on the
phenyl ring. Bis-nitrile 47 resulted in a considerable loss of PI3K potency; possibly due to steric
interference between the bulky isophthalonitrile ring and the enzyme. However, we were pleased to see
that compounds 46, 48 and 49 demonstrated improved isoform selectivity relative to 34. These
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 47
1
2
3
4
experimental results confirmed that small substituents at the 3- and 5-positions on the phenyl ring are well
5
6
accommodated in PI3K.
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
Table 6. Optimization of Phenyl Ring in H2
a
Compound
R3
Isoform Selectivitya, b
PI3KIC50



nM
34
39
40
41
42
43
44
45
46
0.1
2.0
2200
80
4
60
64
2
0.15
0.45
1.9
3100
530
550
140
340
330
230
230
190
90
70
1700
1100
1100
1400
1800
90
0.9
70
3.6
57
1.6
120
210
0.54
47
48
49
150
1.5
>68
850
>68
220
520
>68
220
110
0.47
2600
ACS Paragon Plus Environment

Page 17 of 47
Journal of Medicinal Chemistry
1
2
3
4
aAverage of ≥ 1 determination. Run in quadruplicate. bDetermined from isoform specific PI3K IC50 values.
5
6
7
8
9
While exploring alternate substitutions on the quinazolinone ring, we found that minor modifications had
substantial impact on potency and selectivity. Despite the relatively high clearance of 49 in human
hepatocytes (hHeps = 0.81 L/h/kg) we chose to retain the 3,5-difluoroaniline moiety since it afforded an
enhanced selectivity profile over the unsubstituted phenyl ring of 34. Addition of a fluorine atom to the 6-
position of the quinazolinone, 50, resulted in a 5-fold loss in PI3Kpotency and lower  selectivity
(Table 7). Complete removal of substitutions on the quinazolinone core (51) afforded a less potent and a
more stable compound relative to 49. The diminished hepatic metabolism of 51 may be due to a decrease
in lipophilicity (calculated LogP = 2.6 vs 3.3 (49)). We subsequently focused upon designing compounds
with reduced lipophilicity relative to 49. The fluorine analogs (52-54) were slightly less lipophilic (cLogP
= 2.6 (52-54)) and offered a balance of potency and selectivity. Additionally, we were pleased to find that
both 52 and 53 displayed lower predicted clearances in hHeps in comparison to 49. We saw a substantial
boost in isoform selectivity with the 8-cyano compound (55); unfortunately, 55 exhibited a higher
clearance in hHeps relative to 52. We found 52 to offer us a favorable combination of biochemical
potency, isoform selectivity, and metabolic stability. Additionally, when tested in a whole blood basophil
cellular assay (WBbas), analog 52 potently inhibited PI3K with an EC50 of 1 nM. We further assessed
52 using the KinomeScan platform against a panel of 395 non-mutant kinases at a single concentration of
10 μM. Compound 52 displayed a high degree of kinome selectivity by only binding to other lipid kinases
at less than 10% of control.
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
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 47
1
2
3
4
Table 7. SAR Results for Quinazolinone Ring Modifications
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
a
Compound
R1
Isoform Selectivitya, b
hHeps
L/h/kg
PI3KIC50

2600
780

520
470
290
650
520
920
1200

110
74
nM
0.47
2.9
15
49
50
51
52
53
54
55
5-Cl
5-Cl, 6-F
H
0.81
0.39
0.17
0.34
0.29
0.53
0.5
530
75
5-F
2.2
4.8
3.5
5.9
1900
600
180
160
130
800
6-F
8-F
810
8-CN
1700
aAverage of ≥ 1 determination. Run in quadruplicate. bDetermined from isoform specific PI3K IC50 values.
We next evaluated the PK profile of 52. When administered in vivo, 52 demonstrated low to intermediate
total clearance in rat and dog in comparison to hepatic blood flow in these species, high volumes of
distribution, and high oral bioavailability (Table 8). Rat PK for 52 suggested achievable target coverage at
trough concentrations with twice-daily administration without compromising PI3K isoform selectivity
(Figure 7). Additionally, 52 inhibited ex-vivo anti-IgD stimulation of B-cells with EC50 and EC90 values of
11 and 100 nM, respectively, as determined by a rat whole blood assay. Therefore, 52 was dosed to
maintain these levels in a rat established collagen-induced arthritis (CIA) model.
ACS Paragon Plus Environment

Page 19 of 47
Journal of Medicinal Chemistry
1
2
3
4
Table 8. DMPK Profile of 52
5
6
Compound
Species
Clearance (L/h/kg)
Vss (L/kg) b, c, #
b/pd
t1/2 (h) e, #
%F e, #
in vitroa
1.4
in vivo b, #
0.74 ± 0.04
0.58 ± 0.16
7
8
9
52
Rat
2.6 ± 0.37
1.30 ± 0.1
1.4
2.6
3.2 ± 0.19
106 ± 24
100 ± 40
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
Dog
0.25
2.17 ± 0.56
Human
0.34
aFrom hepatocytes. Intravenous dose of 1 mg/kg. cVolume of distribution at steady state. dBlood plasma ratio. eOral dose of 5
mg/kg. 52 was formulated in 30% PEG 400 and 70% water for dosing in rats for both iv and po; and 5% ethanol, 55% PEG 400,
and 40% citrate buffer at pH 2 (iv) and 15% Captisol (w/v) in water (po) for dosing in dogs. # Number of animals per dosing route
= 3.
b
10000
IV 1 mpk
PO 5mpk
1000
100
10
1
0
6
12
Time (h)
18
24
Figure 7. Concentration-time profiles of 52 in plasma following a 30-minute intravenous infusion at 1 mg/kg and an oral
dose at 5 mg/kg in female Lewis rats (mean ± SD, n=3). The bioavailability of 52 was 106 ± 24%. The Cmax was 2830 ± 942
nM at 1 hr, and AUCinf was 17000 ± 3930 nM·hr.
This experimental model of polyarthritis has been widely used for preclinical testing of numerous anti-
rheumatic agents.20 In this experiment, female Lewis rats with established CIA were treated twice daily
by oral gavage for 7 days starting on the tenth day after collagen treatment, with vehicle or 52 at 0.34,
1.07, 4.6 and 10 mg/kg. Efficacy evaluation was based on reduction in paw swelling—measured by daily
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 47
1
2
3
4
5
6
7
8
9
ankle caliper measurements and terminal hind paw weights—and histopathology parameters
(amelioration of inflammatory infiltrate, pannus formation, cartilage destruction and bone resorption).
Treatment of animals with established CIA with 52 showed significant and dose-dependent reduction in
ankle swelling (Figure 8).
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 8. Compound 52 ameliorates ankle swelling in rat established CIA model. Female Lewis rats with CIA were
treated, twice daily by oral gavage, for 7 days (study days 10–16) with vehicle, 52 (0.34, 1.07, 4.6 or 10 mg/kg), or the
reference compound dexamethasone (0.03 mg/kg). Caliper measurements of right and left ankle diameters were taken
every day beginning on day 9 and continuing through day 16. * p < 0.05 versus arthritis + vehicle.
A correlation between plasma concentration of 52 and the degree of target inhibition leading to efficacy
was established utilizing a rat whole-blood ex-vivo BCR-induced B-cell activation assay. Compound 52
inhibited PI3K-mediated B-cell activation in vivo with an EC50 of 20.5 nM and an EC90 of 100 nM.
These values were used to assess the degree of target inhibition at the dose levels tested in the CIA model.
Plasma concentrations measured on days 3 and 7 were consistent with maintaining approximately 25%,
50%, and greater than 90% inhibition of PI3Kδ at 0.34, 1.07, 4.6 and 10 mg/kg doses, respectively at
trough levels (Figure 9).
ACS Paragon Plus Environment

Page 21 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10 mpk
4.6 mpk
1.07 mpk
0.34 mpk
10000
1000
100
10
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
EC90
EC50
1
0
2
4
6
8
10
12
Time Post Dose (h)
Figure 9. The pharmacokinetic profile of 52 in Lewis rats with established CIA. Blood samples from each dose group
animals at indicated time post dosing on day 3 and 7 were analyzed for plasma concentrations of 52.
Histopathology showed animals treated with 1.07 mg/kg, targeting EC50 inhibition at trough, had
significantly ameliorated ankle inflammation (45% reduction), pannus formation (60%), cartilage damage
(50%), bone resorption (63%), and periosteal bone width (52%), as compared to vehicle controls, with
ED50 values calculated at approximately 1.5 mg/kg on all disease parameters (Figure 10). Maintaining
PI3K coverage between EC50 and EC90 over the dosing interval was sufficient to significantly reduce
disease scores in this rat established model of RA.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 47
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 10. Analog 52 ameliorated histopathological disease severity in rats with established CIA. Ankle and knee joint
tissue, sections from all animals were examined microscopically and graded on a scale from 0 to 5 for each parameter
based on the degree of severity, * p < 0.05 versus arthritis + vehicle. (See supporting information for scoring guidelines).
CONCLUSION
In summary, we have optimized a series of potent and highly selective PI3K inhibitors. During the
course of our SAR we discovered a novel triaminopyrimidine hinge binder that forms three hydrogen
bonds with the protein. We also optimized the phenyl and the quinazolinone regions for potency and
isoform selectivity. Compound 52, although not suitable for once-daily dosing, was identified as a potent
and selective inhibitor of PI3K with good pharmacokinetic properties. Moreover, 52 demonstrated
efficacy in a model of rat CIA with less than complete target coverage over the dosing interval. Further
work and additional optimization to achieve a low dose PI3K inhibitor suitable for QD administration
will be presented in a future publication.
EXPERIMENTAL SECTION
General Procedures. All final compounds were determined to be at least 95% pure by 1H NMR and
LCMS analysis. All commercial reagents and anhydrous solvents were used as provided without further
ACS Paragon Plus Environment

Page 23 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
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. For all samples 0.1% TFA was added to both eluents.
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)-2,4-diamino-6-((1-(5-chloro-4-oxo-3-phenyl-3,4-dihydroquinazolin-2-yl)ethyl)amino)pyrimidine-
5-carbonitrile 2,2,2-trifluoroacetic acid salt (34). (Synthesized according to the procedure described in
Scheme 1). 2,4-diamino-6-chloropyrimidine-5-carbonitrile (391 mg, 2.31 mmol)was added to a
suspension of (S)-2-(1-aminoethyl)-5-chloro-3-phenylquinazolin-4(3H)-one 2,2,2-trifluoroacetic acid salt
(865 mg, 2.89 mmol) (7) and diisopropylethylamine (7.6 mL, 43.3 mmol) in IPA (3 mL). The resultant
mixture was heated to 130oC for 14 hours, after which time the reaction was cooled to room temperature.
The resultant solid was filtered, and purified by HPLC eluting with 5%-95% water/acetonitrile (0.1%v/v
trifluoroacetic acid). The appropriate fractions were pooled and lyophilized to afford the title compound
1
as a white solid (575 mg). H NMR (400 MHz, DMSO) δ 7.78 (t, J = 8.0 Hz, 1H), 7.63 (dd, J = 8.2, 1.2
Hz, 1H), 7.57 (dd, J = 7.8, 1.2 Hz, 1H), 7.57 – 7.45 (m, 5H), 6.79 (d, J = 7.0 Hz, 1H), 6.55 (br s, 1H), 6.25
(br s, 1H), 4.72 (dq, J = 6.45, 6.45 Hz, 1H), 1.29 (d, J = 6.8 Hz, 3H). ES/MS m/z 433.1 (M+H) +
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 47
1
2
3
4
5
6
7
8
9
(S)-2-(1-aminoethyl)-5-chloro-3-phenylquinazolin-4(3H)-one 2,2,2-trifluoroacetic acid salt.
Trifluoroacetic acid (3 mL) was added to a solution of (S)-tert-butyl 1-(5-chloro-4-oxo-3-phenyl-3,4-
dihydroquinazolin-2-yl)ethylcarbamate (1 g, 2.5 mmol) in dichloromethane (3 mL). The resultant was
stirred at room temperature for 3 hours. The solvents were removed in vacuuo to afford the title
compound as a white solid (823 mg). ES/MS m/z = 300.1 (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)-tert-butyl 1-(5-chloro-4-oxo-3-phenyl-3,4-dihydroquinazolin-2-yl)ethylcarbamate. Diphenyl
phosphite (1.9 mL, 10 mmol) was added to a solution of 2-amino-6-chlorobenzoic acid (495 mg, 2.9
mmol) (2) and (S)-2-(tert-butoxycarbonylamino)propanoic acid (710 mg, 3.77 mmol) in pyridine (3 mL).
The reaction mixture was stirred at 40 oC for 2 hours. Aniline (274 mg, 3.48 mmol) was then added to the
reaction mixture, which was then stirred at 55 oC for 12 hours. The reaction mixture was cooled to room
temperature. This mixture was then diluted with EtOAc (50 mL), washed with 1N aqueous HCl (2 x
50mL), brine (50mL), and dried over sodium sulfate. The organics layer was filtered and concentrated in
vacuuo to afford material which was purified by column chromatography on SiO2 eluting with EtOAc in
hexanes (0-50%) to afford the title compound as a solid (1 g). ES/MS m/z = 400.1 (M+H) +
2,4-diamino-6-chloropyrimidine-5-carbonitrile. 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 oC and stirred for 3 hrs. The reaction mixture was cooled to
10oC and water (50 mL) was added. 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) +
(S)-2-(1-((6-amino-5-((5-fluoropyridin-2-yl)ethynyl)pyrimidin-4-yl)amino)ethyl)-5-chloro-3-
phenylquinazolin-4(3H)-one 2,2,2-trifluoroacetate (31). (Synthesized according to the procedure
described in Scheme 2). To (S)-2-(1-aminoethyl)-5-chloro-3-phenylquinazolin-4(3H)-one 2,2,2-
ACS Paragon Plus Environment

Page 25 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
trifluoroacetic acid salt. (58 mg, 0.14 mmol) and 6-chloro-5-((5-fluoropyridin-2-yl)ethynyl)pyrimidin-4-
amine (35 mg, 0.14 mmol) in IPA (2 mL) was added Hünig’s base (0.14 mL, 0.7 mmol). The reagents
were stirred at 130oC for 14 hours in a microwave reactor after which time the reaction mixture was
cooled to room temperature. The solvent was evaporated and resultant solid was purified by HPLC
eluting with 5%-95% water/acetonitrile (0.1%v/v trifluoroacetic acid). The appropriate fractions were
pooled and lyophilized to afford the title compound as a solid (32 mg) 1H NMR (400 MHz, DMSO-d6) δ
8.66 (d, J = 3.0 Hz, 1H), 7.88 (dd, J = 8.0, 4.0 Hz, 1H), 7.85 – 7.71 (m, 4H), 7.64 – 7.35 (m, 6H), 7.11 (d,
J = 6.9 Hz, 1H), 6.83 (s, 2H), 4.67 (p, J = 6.5 Hz, 1H), 1.24 (d, J = 6.6 Hz, 3H). ES/MS m/z = 512.1
(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
6-chloro-5-((5-fluoropyridin-2-yl)ethynyl)pyrimidin-4-amine. To 6-chloro-5-iodopyrimidin-4-amine
(150 mg, 0.58 mmol) ( in DMF (2 mL) was added Pd(PPh3)2Cl2 (28 mg, 0.04 mmol), CuI (5.5 mg, 0.029
mmol), TEA (2 mL), and 2-ethynyl-5-fluoropyridine (70 mg, 0.58 mmol). The reaction mixture was
stirred at 90 oC for 2 hours, after which time the reaction mixture was cooled to ambient temperature and
partitioned between Ethyl acetate (20 mL) and Water (20 mL). The organic layer was washed with brine
(25 mL), separated, dried over magnesium sulfate and filtered. The resultant residue was purified on silica
with 0 to 10% Methanol in Dichloromethane to afford the title compound as a solid (80 mg). ES/MS m/z
= 249.3 (M+H) +
2-ethynyl-5-fluoropyridine. To 5-fluoro-2-((trimethylsilyl)ethynyl)pyridine (400 mg, 2.06 mmol) in
THF (5 mL) was added Tetra-n-butylammonium fluoride (2.0 mL, 2.0 mmol). The reaction mixture was
stirred at ambient temperature for 3 hours. After which time the solvent was evaporated and the resultant
residue was purified on silica gel with 0 to 100% Ethyl Acetate in Hexanes to afford the title compound
as an oil (150 mg). ES/MS m/z = 122.1 (M+H) +
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 47
1
2
3
4
5
6
7
8
9
5-fluoro-2-((trimethylsilyl)ethynyl)pyridine. To 2-bromo-5-fluoropyridine (500mg, 2.84 mmol) in THF
(5 mL) was added Pd(PPh3)2Cl2 (136 mg, 0.19 mmol), CuI (26 mg, 0.13 mmol), TEA (5 mL), and
ethynyltrimethylsilane (1.9 ml, 13.9 mmol). The reaction mixture was stirred at 70oC for 2 hours, after
which time the solvent was evaporated. The resultant residue was purified on silica gel with 0 to 100%
Ethyl Acetate in Hexanes to afford the title compound as a solid (500 mg). ES/MS m/z = 226.3 (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-(1-((6-aminopyrimidin-4-yl)amino)ethyl)-5-chloro-3-phenylquinazolin-4(3H)-one (15). The
target compound was prepared in a manner analogous to that described for compound 34 as shown in
1
Scheme 1. H NMR (400 MHz, CDCl3) δ 8.1 (s, 1), 7.7 – 7.3 (m, 8H), 5.63 (m, 1H), 5.35 (br s, 1H), 4.8
(m, 1H), 4.70 (br, 2H), 1.38 (d, 3H). ES/MS m/z 393.1 (M+H) +
(S)-5-chloro-3-phenyl-2-(1-(pyrimidin-4-ylamino)ethyl)quinazolin-4(3H)-one (16). The target
compound was prepared in a manner analogous to that described for compound 34 as shown in Scheme 1.
1H NMR (400 MHz, DMSO-d6) δ 9.85 (d, 1H), 8.75 (s, 1H), 8.2, (m, 1H), 7.8 (m, 1H), 7.7 – 7.4 (m, 7H),
6.8 (d, 1H), 4.70 (m, 1H), 1.4 (d, 3H). ES/MS m/z 378.1 (M+H) +
(S)-4-((1-(5-chloro-4-oxo-3-phenyl-3,4-dihydroquinazolin-2-yl)ethyl)amino)pyrimidine-5-
carbonitrile (17). The target compound was prepared in a manner analogous to that described for
1
compound 34 as shown in Scheme 1. H NMR 1H NMR (400 MHz, DMSO-d6) δ 8.60 (bs, 1H), 8.51 –
8.46 (m, 2H), 7.83 – 7.74 (m, 1H), 7.66 (dd, J = 8.2, 1.2 Hz, 1H), 7.59 (dd, J = 7.8, 1.2 Hz, 1H), 7.57 –
7.52 (m, 1H), 7.52 – 7.46 (m, 1H), 7.46 – 7.41 (m, 1H), 7.40 – 7.30 (m, 2H), 4.90 (q, J = 6.7 Hz, 1H),
1.43 (d, J = 6.7 Hz, 3H). ES/MS m/z 403.1 (M+H) +
(S)-4-amino-6-((1-(5-chloro-4-oxo-3-phenyl-3,4-dihydroquinazolin-2-yl)ethyl)amino)pyrimidine-5-
carbonitrile (18) as shown in Scheme 1. The target compound was prepared in a manner analogous to
ACS Paragon Plus Environment

Page 27 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
that described for compound 34. 1H NMR (400 MHz, DMSO-d6) δ 7.85 (s, 1H), 7.8 – 7.4 (m, 8H), 7.25
(br, 2H), 4.75 (m, 1H), 1.33 (d, 3H). ES/MS m/z 418.1 (M+H) +
6
7
8
9
(S)-2-amino-4-((1-(5-chloro-4-oxo-3-phenyl-3,4-dihydroquinazolin-2-yl)ethyl)amino)nicotinonitrile
(19). The target compound was prepared in a manner analogous to that described for compound 34 as
shown in Scheme 1. 1H NMR (400 MHz, DMSO-d6) δ 7.80 (d, J = 7.9 Hz, 1H), 7.67 (d, J = 8.0 Hz, 1H),
7.60 (d, J = 7.7 Hz, 1H), 7.54 (s, 2H), 7.49 – 7.37 (m, 5H), 6.07 (s, 1H), 4.78 – 4.59 (m, 1H), 1.38 (s, 3H).
ES/MS m/z 417.2 (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-((1-(5-chloro-4-oxo-3-phenyl-3,4-dihydroquinazolin-2-yl)ethyl)amino)pyrimidine-5-
carbonitrile (20) The target compound was prepared in a manner analogous to that described for
1
compound 34 as shown in Scheme 1. H NMR (400 MHz, DMSO-d6) δ 8.12 (s, 1H), 7.8 (d, 1H), 7.64
(dd, 1H), 7.58 (d, 1H), 7.36-7.56 (m, 5H), 6.9 (br, 3H), 4.8 (m, 1H), 1.35 (d, 3H). ES/MS m/z 418.1
(M+H) +
(S)-4-amino-6-((1-(5-chloro-4-oxo-3-phenyl-3,4-dihydroquinazolin-2-yl)ethyl)amino)-2-
hydroxypyrimidine-5-carbonitrile (21). To (S)-4-amino-6-((1-(5-chloro-4-oxo-3-phenyl-3,4-
dihydroquinazolin-2-yl)ethyl)amino)-2-(methylsulfonyl)pyrimidine-5-carbonitrile (prepared as described
for 33) in dioxane was added 1M aqueous NaOH (10 Eq). The reaction mixture was allowed to stir at RT
for 2 hours. LCMS indicated consumption of starting material showing mostly desired sulfone and some
sulfoxide. The material was extracted into EtOAc, concentrated to a residue which was taken up in
DMSO and purified by RF HPLC. Clean fractions were pooled and lyophilized to give the desired
product as a white powder (36% yield). 1H NMR (400 MHz, DMSO-d6) δ 7.82 – 7.75 (m, 1H), 7.75 –
7.69 (m, 2H), 7.63 (dd, J = 8.2, 1.2 Hz, 1H), 7.58 (dd, J = 7.8, 1.2 Hz, 1H), 7.56 – 7.33 (m, 3H), 4.62 (t, J
= 6.7 Hz, 1H), 1.33 (d, J = 6.7 Hz, 3H). ES/MS m/z 434.1 (M+H) +
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 47
1
2
3
4
5
6
7
8
9
(S)-4-amino-6-((1-(5-chloro-4-oxo-3-phenyl-3,4-dihydroquinazolin-2-yl)ethyl)amino)pyrimidine-5-
carboxamide (22). The target compound was prepared in a manner analogous to that described for
compound 34 as shown in Scheme 1. 1H NMR (400 MHz, DMSO-d6) δ 7.88 (d, 1H), 7.83 (s, 1H), 7.75 (t,
1H), 7.4-7.6 (m, 8H), 6.6 (br, 2H), 4.7 (m, 1H), 4.1 (br 1H), 3.3 (br 1H), 1.25 (d, 3H). ES/MS m/z 436.1
(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-(1-((6-amino-5-(trifluoromethyl)pyrimidin-4-yl)amino)ethyl)-5-chloro-3-phenylquinazolin-
4(3H)-one (23). The target compound was prepared in a manner analogous to that described for
compound 34 as shown in Scheme 1. 1H NMR (400 MHz, DMSO-d6) δ 7.91 (s, 1H), 7.77 (t, 1H), 7.4-7.7
(m, 7H), 6.8 (br, 3H), 4.8 (m, 1H), 1.3 (d, 3H). ES/MS m/z 461.1 (M+H) +
(S)-2-(1-((6-amino-5-methylpyrimidin-4-yl)amino)ethyl)-5-chloro-3-phenylquinazolin-4(3H)-one
(24). The target compound was prepared in a manner analogous to that described for compound 34 as
shown in Scheme 1. 1H NMR (400 MHz, DMSO-d6) δ 7.4-7.8 (m, 9H), 6.3 (d, 1H), 5.87 (br, 2H), 4.56
(m, 1H), 1.84 (s, 3H), 1.34 (d, 3H). ES/MS m/z 407.1(M+H) +
(S)-2-(1-((6-amino-5-methoxypyrimidin-4-yl)amino)ethyl)-5-chloro-3-phenylquinazolin-4(3H)-one
(25). The target compound was prepared in a manner analogous to that described for compound 34 as
shown in Scheme 1. 1H NMR (400 MHz, DMSO-d6) δ 8.00 (s, 1H), 7.77 (m, 1H), 7.65 – 7.54 (m, 2H),
7.52 – 7.43 (m, 2H), 7.40 (s, 3H), 4.73 (m, 1H), 3.65 (s, 3H), 1.49 – 1.31 (m, 3H). ES/MS m/z 423.1
(M+H) +
(S)-2-(1-((6-amino-5-ethynylpyrimidin-4-yl)amino)ethyl)-5-chloro-3-phenylquinazolin-4(3H)-one
(26). The target compound was prepared in a manner analogous to that described for compound 31 as
shown in Scheme 2, using 12 and Ethynyltrimethylsilane as the starting materials. 1H NMR (400 MHz,
ACS Paragon Plus Environment

Page 29 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
DMSO-d6) δ 7.82 (s, 1H), 7.76 (t, 1H), 7.4-7.7 (m, 7H), 6.7 (d, 1H), 6.54 (br, 2H), 4.77 (s, 1H), 4.7 (m,
1H), 1.3 (d, 3H). ES/MS m/z 417.1 (M+H) +
7
8
9
(S)-2-(1-(6-amino-5-(phenylethynyl)pyrimidin-4-ylamino)ethyl)-5-chloro-3-phenylquinazolin-4(3H)-
one (27). The target compound was prepared in a manner analogous to that described for compound 31 as
shown in Scheme 2. 1H NMR (400 MHz, DMSO-d6) δ 7.88 (s,1H), 7.78 (t, J = 8.0 Hz, 1H), 7.88 (dt, J =
6.8, 1.6 Hz, 2H), 7.59 – 7.41 (m, 10H), 6.92 (d, J = 7.0 Hz, 1H), 6.75 (br, 2H), 4.73 (p, J = 6.8 Hz, 1H),
1.35 (d, J = 6.7 Hz, 3H). ES/MS m/z 493.1 (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-(1-(6-amino-5-(pyridin-3-ylethynyl)pyrimidin-4-ylamino)ethyl)-5-chloro-3-phenylquinazolin-
4(3H)-one (28). The target compound was prepared in a manner analogous to that described for
1
compound 31 as shown in Scheme 2. H NMR (400 MHz, DMSO-d6) δ 8.96 (d, J = 1.3 Hz, 1H), 8.62
(dd, J = 4.9, 1.7 Hz, 1H), 8.30 – 8.03 (m, 2H), 7.98 (s, 1H), 7.87 – 7.69 (m, 2H), 7.66 – 7.30 (m, 6H),
4.93 – 4.69 (m, 1H), 1.44 (d, J = 6.7 Hz, 3H)). ES/MS m/z 494.1 (M+H) +
(S)-2-(1-(6-amino-5-(pyrazin-2-ylethynyl)pyrimidin-4-ylamino)ethyl)-5-chloro-3-phenylquinazolin-
4(3H)-one (29). The target compound was prepared in a manner analogous to that described for
compound 31 as shown in Scheme 2. 1H NMR (400 MHz, DMSO-d6) δ 9.03 (d, J = 1.6 Hz, 1H), 8.78
(dd, J = 2.5, 1.6 Hz, 1H), 8.60 (d, J = 24.1 Hz, 1H), 8.07 (s, 1H), 7.92 – 7.68 (m, 2H), 7.64 – 7.23 (m,
9H), 4.89 – 4.61 (m, 1H), 1.34 (d, J = 6.7 Hz, 3H). ES/MS m/z 495.4 (M+H) +
(S)-2-(1-(6-amino-5-(pyridazin-3-ylethynyl)pyrimidin-4-ylamino)ethyl)-5-chloro-3-
phenylquinazolin-4(3H)-one (30). The target compound was prepared in a manner analogous to that
described for compound 31 as shown in Scheme 2. 1H NMR (400 MHz, DMSO-d6) δ 8.73 (d, J = 2.9 Hz,
1H), 8.01 – 7.79 (m, 4H), 7.73 – 7.39 (m, 8H), 7.38 – 7.17 (m, 2H), 4.85 – 4.74 (m, 1H), 1.35 (dd, J =
13.1, 6.6 Hz, 3H). ES/MS m/z 496.1 (M+H) +
ACS Paragon Plus Environment