Discovery of AB680: A Potent and Selective Inhibitor of CD73

Article abstract of DOI:10.1021/acs.jmedchem.0c00525

Extracellular adenosine (ADO), present in high concentrations in the tumor microenvironment (TME), suppresses immune function via inhibition of T cell and NK cell activation. Intratumoral generation of ADO depends on the sequential catabolism of ATP by two ecto-nucleotidases, CD39 (ATP → AMP) and CD73 (AMP → ADO). Inhibition of CD73 eliminates a major pathway of ADO production in the TME and can reverse ADO-mediated immune suppression. Extensive interrogation of structure-activity relationships (SARs), structure-based drug design, and optimization of pharmacokinetic properties culminated in the discovery of AB680, a highly potent (Ki = 5 pM), reversible, and selective inhibitor of CD73. AB680 is further characterized by very low clearance and long half-lives across preclinical species, resulting in a PK profile suitable for long-acting parenteral administration. AB680 is currently being evaluated in phase 1 clinical trials. Initial data show AB680 is well tolerated and exhibits a pharmacokinetic profile suitable for biweekly (Q2W) iv-administration in human.

Full text of DOI:10.1021/acs.jmedchem.0c00525

Subscriber access provided by UCL Library Services
Drug Annotation
Discovery of AB680 – A Potent and Selective Inhibitor of CD73
Kenneth V. Lawson, Jaroslaw Kalisiak, Erick A Lindsey, Eric Newcomb, Manmohan Reddy Leleti, Laurent
Debien, Brandon R. Rosen, Dillon H Miles, Ehesan U. Sharif, Jenna Jeffrey, Joanne B. L. Tan, Ada Chen,
Sharon Zhao, Guifen Xu, Lijuan Fu, Lixia Jin, Tim W Park, Wade Berry, Susanne Moschütz, Emma R Scaletti,
Norbert Sträter, Nigel P Walker, Stephen W Young, Matthew J Walters, Uli Schindler, and Jay P Powers
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.0c00525 • Publication Date (Web): 02 Jul 2020
Downloaded from pubs.acs.org on July 2, 2020
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 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 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.
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 46
Journal of Medicinal Chemistry
Discovery of AB680 – A Potent and Selective Inhibitor of CD73
1
2
3
4
5
6
7
8
9
Kenneth V. Lawson#*, Jaroslaw Kalisiak#, Erick A. Lindsey#, Eric T. Newcomb#, Manmohan Reddy Leleti#, Laurent Debien#,
Brandon R. Rosen#, Dillon H. Miles#, Ehesan U. Sharif#, Jenna L. Jeffrey#, Joanne B.L. Tan#, Ada Chen#, Sharon Zhao#,
Guifen Xu#, Lijuan Fu, Lixia Jin#, Tim W. Park#, Wade Berry#, Susanne Moschütz±, Emma Scaletti±, Norbert Sträter±, Nigel
P. Walker#, Stephen W. Young#, Matthew J. Walters#, Uli Schindler#, Jay P. Powers#
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
#Arcus Biosciences, Inc., 3928 Point Eden Way, Hayward, California 94545, United States
±Institute of Bioanalytical Chemistry, Center for Biotechnology and Biomedicine, Leipzig University, Deutscher Platz 5,
D-04103 Leipzig, Germany.
Abstract:
Extracellular adenosine (ADO), present in high concentrations in the tumor microenvironment (TME), suppresses immune
function via inhibition of T cell and NK cell activation. Intra-tumoral generation of ADO depends on the sequential
catabolism of ATP by two ecto-nucleotidases, CD39 (ATP→AMP) and CD73 (AMP→ADO). Inhibition of CD73 eliminates a
major pathway of ADO production in the TME and can reverse ADO-mediated immune suppression. Extensive
interrogation of structure activity relationships (SAR), structure-based drug design, and optimization of pharmacokinetic
properties culminated in the discovery of AB680, a highly potent (Ki = 5 pM), reversible, and selective inhibitor of CD73.
AB680 is further characterized by very low clearance and long half-lives across preclinical species, resulting in a PK profile
suitable for long-acting parenteral administration. AB680 is currently being evaluated in Phase 1 clinical trials. Initial data
show AB680 is well tolerated and exhibits a pharmacokinetic profile suitable for biweekly (Q2W) IV-administration in
human.
Introduction.
The introduction of monoclonal antibody (mAb) therapies acting as immune checkpoint inhibitors, such as those targeting
programmed cell death 1 (PD-1) and cytotoxic T-lymphocyte antigen 4 (CTLA-4), has revolutionized cancer immunotherapy
treatment.1,2 However, these agents only elicit strong responses in a minority of patients.3 For non-responders, tumors
exploit additional immunosuppressive mechanisms to facilitate immune escape.4 As such, combination therapies designed
to block multiple immunosuppressive pathways may be required to improve clinical responses.5 One mechanism by which
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 46
tumors evade the immune system is via depletion or increased production of small molecule signaling compounds within
the tumor microenvironment (TME). Falling into the latter category, adenosine (ADO) has been shown to accumulate in
tumors and exert profound immunosuppressive effects on a variety of tumor-infiltrating leukocytes (TILs) via the A2A/A2B
adenosine receptors.6 Intra-tumoral production of ADO is regulated by the ecto-nucleotidase CD73, which converts
extracellular adenosine monophosphate (AMP) to ADO.7,8,9,10 CD73 is expressed on a variety of TILs and has been found
to be overexpressed on a variety of human cancers. Furthermore, high CD73 expression has been associated with poor
clinical prognosis.11 Inhibition of CD73 is expected to reverse ADO-mediated immune suppression by blocking the major
mechanism for ADO production in the TME.12,13,14,15
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
CD73 is a glycosylphosphatidylinositol (GPI-) anchored homodimeric ecto-nucleotidase, that undergoes an
extensive conformational switch between open and closed forms upon substrate binding and release.16 The binding mode
of nucleotide-derived inhibitor adenosine 5’-(α,β-methylene)diphosphate (AMPCP, 1, Ki = 88.4 nM, Figure 1) has been
elucidated and proven instrumental in the design of nucleoside derived CD73 inhibitors.16,17 Müller and co-workers
reported the initial structure-activity relationships (SAR) of this series, which resulted in the discovery of PSB-12379 (2, Ki
= 2.2 nM, Figure 1),18,19,26 and, following completion of our CD73 inhibitor discovery effort20,21, have reported optimized
CD73 inhibitors based on that scaffold.22,23
A. CD73 enzymatic activity
NH2
N
NH2
N
N
N
N
N
O
P
N
N
O
O
CD73
HO
O
HO
OH
HO
HO
OH
OH
PO4
adenosine-
monophosphate
(AMP)
adenosine
(ADO)
B. Known nucleotide CD73 inhibitors
R
1
AMPCP (R = H)
HN
6
hCD73 Ki = 88.4 nM
hCD73 IC50 = 187 nM
5
N
1
N
O
P
O
P
2
5'
4
1' N
N
O
4'
3'
3
2
PSB-12379 (R = Bn)
HO
O
OH OH
hCD73 Ki = 2.2 nM
2'
hCD73 IC50 = 4.3 nM
HO
OH
Figure 1. A. CD73 catalyzes the conversion of AMP to ADO
with loss of inorganic phosphate (Pi). B. Structure of known
CD73 inhibitors AMPCP and PSB-12379
ACS Paragon Plus Environment

Page 3 of 46
Journal of Medicinal Chemistry
AMPCP is a competitive inhibitor that binds the active site of CD73 via a concert of ionic, polar and hydrophobic
1
2
interactions.16 The ribose moiety forms bifurcated hydrogen bonds with D506 and N390, while the adenine heterocycle is
sandwiched between F417 and F500 in a face-to-face π-stacking interaction (Figure 2A). Lastly, the bisphosphonate group
forms an ionic interaction with the di-Zn catalytic domain with an additional ion-pair with R354 (Figure 2B). Our early
investigations revealed modification or replacement of the bisphosphonate moiety was poorly tolerated within the
nucleotide-mimetic scaffold. This observation confounded our goal of developing an orally bioavailable CD73 inhibitor as
phosphate-containing small molecules are notorious for their poor oral bioavailability. Furthermore, high throughput and
in silico screening techniques generated few or low-quality leads for further optimization. With these revelations we were
faced with the option of guiding our discovery effort towards alternative routes of administration or rely upon pro-drug
strategies.24 The latter challenge appeared insurmountable due to the likely necessity to mask three ionizable positions
and the sparsity of relevant precedent to do so.
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
We undertook medicinal chemistry efforts in pursuit of a molecule suitable for IV-administration at extended
dosing intervals. For the convenience of patients, our goal was to identify a candidate which could be administered
biweekly (Q2W), concomitant with typical mAb checkpoint inhibitor dosing schedules.25 Regardless of the degree of
potency we would ultimately be able to achieve, we realized that extraordinarily low clearance would be a necessity to
continually inhibit CD73 activity over the desired dosing interval. As such, our discovery effort relied heavily on the
optimization of pharmacokinetic parameters and the early development of tractable structure-pharmacokinetic
relationships to ascertain the feasibility of this approach.
Results and Discussion. Our initial investigations into the SAR of the adenosine scaffold was focused on modification of
the ribose moiety to avoid the known pharmacokinetic liabilities of the AMPCP scaffold. This work was performed with
the 2-chloroadenine base, as in 3, as this has been shown to improve potency.26 In our assay, we found that incorporation
of chlorine at the C2-position (3, Figure 2C) of adenine resulted in a 35-fold improvement in potency relative to AMPCP.
With both the 2’ and 3’ hydroxyl groups of 3 putatively engaged in hydrogen bonds with CD73, we were pleased to find
that replacement of the 2’ hydroxyl group with fluorine was well tolerated. The fluorine stereochemistry appeared to be
important as incorporation of 2’-(R)-fluoro in earlier scaffolds (5) had resulted in substantial loss in potency.
.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 46
A.
B.
H
N
1
2
3
4
5
6
7
8
F500
H220
D85
O
O
NH2
N
F417
O
His
O
D36
H38
N
H243
N
O
O
O
O
O
Zn
Zn
O
Zn
Zn
O
N
P
N
N
HO
HO
N
N
O
O
O
O
-stack

NH
P
O P
O
O
H
H
H
O
H
O
N
D506
O
P
O
O
'top'
view
'side'
view
O
H-bond donor/acceptor
N390
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
C. Initial structure-activity relationship investigation
NH2
NH2
NH2
N
N
N
N
N
N
N
N
O
P
O
P
O
P
O
P
O
P
O
P
N
Cl
N
N
Cl
N
O
O
O
HO
O
HO
O
HO
O
OH OH
OH OH
OH OH
HO
HO
HO
OH
F
F
(3)
(4)
2'-(S)-F, 2-Cl-AMPCP
2-Cl-AMPCP
(5)
2'-(R)-F-AMPCP
hCD73-CHO IC50 = 5.0 nM
hCD73-CHO IC50 = 3.4 nM
hCD73-CHO IC50 = 18 M
Figure 2. A. Schematic of key binding interactions between AMPCP and CD73 (closed form) based on PDB 4H2I
which guided structure-based design and initial SAR. B. Schematic of ionic interaction of bisphosphonate
moiety and di-Zn catalytic site. C. Initial SAR investigation.
Work by Müller and co-workers demonstrated that considerable potency gains could be achieved by incorporation
of hydrophobic groups at the C6-position.18,26 Our efforts identified a similar trend within the β-fluoroarabinose scaffold
and we prepared numerous ribose examples for direct comparison. Although incorporation of small motifs, such as
methyl-, ethyl-, or cyclopropylamine (Table 1, entries 6-9) did not improve potency, the larger cyclopentylamine group
(Table 1, entries 12-13) resulted in a 10-fold potency improvement, providing subnanomolar CD73 inhibitors in both the
ribose and β-fluoroarabinose scaffolds. Substitution with benzylamine (entries 16-17) also resulted in similarly improved
potency. Beyond these initial discoveries, the SAR at this position was relatively flat. Extensive elaboration of the
benzylamine analogs was performed (e.g. 22-24), the majority of which resulted in equivalent or slightly reduced potency.
Increasing the polarity of the C6-substituent (entries 26-27) resulted in moderate loss in potency, and disubstituted amines
(e.g. pyrrolidine, 25) were well tolerated, indicating that the N-H was not involved in a hydrogen bond donor interaction.
A similar conclusion was made by the Muller lab working on a related scaffold.18
ACS Paragon Plus Environment

Page 5 of 46
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
IC50
IC50
Entry
10
R1
R2
Entry
R1
R2
R1
N
(nM)
(nM)
N
N
22
23
24
(o) 4.2
(m) 1.3
(p) 2.4
N
(S)-F
3.6
0.70
0.40
0.20
0.15
0.12
0.70
0.30
1.6
Me
Me
Cl
O
P
O
P
HN
(S)-F
Cl
O
HN
HN
11
(R)-OH
(S)-F
O
HO
OH OH
8
9
R2
HO
12
25
26
27
28
29
(S)-F
(S)-F
(S)-F
(S)-F
(S)-F
0.90
0.80
3.6
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
IC50
(nM)
Entry
R1
R2
13
(R)-OH
(S)-F
3a
4
(R)-OH
(S)-F
5.0
3.4
14
NH2
Me
O
N
HN
15
(R)-OH
(S)-F
16
Me
HN
HN
HN
6
7
(S)-F
(S)-F
12.0
5.0
17a
18
(R)-OH
(S)-F
N
Me
1.1
HN
Me
HN
HN
N
19a
20
(R)-OH
(S)-F
0.70
4.1
Me
8
9
(S)-F
2.6
1.7
O
1.3
HN
(R)-OH
21a
(R)-OH
0.30
Table 1. Modification of C6 position (R1) in ribose[R2 = (R)-OH] and β-fluoroarabinose[R2 = (S)-F] scaffolds. aReference 26.
Completion of this study revealed a consistent trend between the two furanose scaffolds, in which the β-
fluoroarabinose scaffold generally demonstrated only slightly reduced potency relative to equivalently functionalized
ribose analogs. This came as a surprise given that the 2’-hydroxyl group of AMPCP is reported to form two hydrogen bond
interactions with D506 and N390 that would presumably be absent in β-fluoroarabinose analogs.9 Furthermore, despite
the modest loss in potency, the β-fluoroarabinose scaffold offered a compelling program direction, as compounds in this
series generally exhibited improved physicochemical and pharmacokinetic properties. As such, we sought to further
understand the molecular interactions of the β-fluoroarabinose core with the CD73 catalytic site. To accomplish this, we
obtained a high-resolution co-crystal structure of 16 bound to the closed form of CD73.
The key interactions of 16 with CD73 are depicted in Figure 3A. Consistent with the binding interactions of AMPCP,
the bisphosphonate group binds the di-zinc catalytic site and the purine heterocycle is sandwiched between F417 and
F500 in a π-stacking interaction. In contrast, key interactions of the β-fluoroarabinose moiety were unique. The arabinose
ring exhibited a unique puckered conformation moving the 3’-hydroxyl group into close proximity with N390. This
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 46
conformation allowed for a similar bifurcated hydrogen bond between the side-chain of N390 and both the 3’-OH and
adenine-N3 (Figure 3A). The consequence of this altered ring conformation is best displayed by the overlay of AMPCP with
16 (Figure 3B). The 3’-OH of 16 is located roughly equidistant between the 2’-OH and 3’-OH of AMPCP. This provides a
strong rationale for the retained potency of this series, which lacks a hydrogen bond acceptor at the 2’-position. If the
ribose conformation of 16 was conserved with that of AMPCP, a hydrogen bond with N390 would be unattainable.
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
A
B
Asp-085
His-038
Asp-036
Phe-417
His-220
His-243
Asn-245
Phe-500
Arg-354
Asp-506
Asp-506
Asn-390
Asn-390
D
C
Figure 3. A. Key interactions between 16 and CD73 (closed form), 1.94 Å resolution, PDB ID: 6Z9B. B. Overlay of
16 (yellow) and AMPCP (pink, PDB ID: 4H2I). Key interaction with N390 and D506 shown. (Superposition based
on matched atom pairs) C. C2-pocket cavity surface (grey mesh) is lined predominantly by hydrophobic residues.
D. Changes in the water structure of the C2-pocket in a comparison of the binding modes of AMPCP (ligand,
protein residues and water molecules in grey) and 16 (ligand yellow, protein residues blue-purple, waters red).
ACS Paragon Plus Environment

Page 7 of 46
Journal of Medicinal Chemistry
The adenine-C2-chlorine is positioned at the entry to a large pocket, occupied by water molecules and lined
1
2
primarily by hydrophobic residues (grey mesh, Figure 3C). The substantial potency enhancement resulting from chlorine
incorporation may be due in part to Van der Waals forces with hydrophobic residues lining the entry to this pocket, polar
interactions with N499, and rearrangements of the C2-pocket water structure. In a comparison of the CD73 cocrystal
structures with AMPCP and 16, introduction of the C2 chlorine substituent in 16 displaces two water molecules (D and E
in Figure 4D) and two new water positions (J and G) are occupied instead, close to the hydrophobic bottom of the pocket.
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
We believed further potency gain could be achieved by occupying a larger fraction of the pocket. We prepared a
number of analogs to interrogate this hypothesis. Unfortunately, it became evident that increasing the size of substituents
at the C2 position was poorly tolerated (Table 2). This may be explained by the narrow entrance to the hydrophobic cavity
imparted by the phenylalanine residues. Increasing the size of the C2 substituent may, in turn, partially disrupt the
favorable π-stacking interaction. Indeed, increasing size from -methyl (30, IC50 = 1.3 nM) to -ethyl (31, IC50 = 12.0 nM)
resulted in a nearly 10-fold potency loss while a C2-ethyne substituent, featuring a smaller steric footprint, was
comparatively well-tolerated (39, IC50 = 2.0 nM). Electron-donating groups -NHMe, -pyrrolidine, and -OMe (34-36,
respectively) were also poorly tolerated (7 to 160-fold loss of potency). These comparatively electron-rich heterocycles
may weaken the π-stacking interaction with F417 and F500. A similar steric argument could also account for their
diminished potency.
Interestingly, a benzyl group at the C2-position (38, IC50 = 1.5 nM) was comparatively well tolerated considering
its increased size. The arene likely forms favorable interactions with residues deeper in the C2-pocket and would be
expected to displace additional water from the pocket. We believe this may offset the aforementioned unfavorable
interactions of a hydrophobic group with the narrow entrance to the C2-pocket. Increasing the size to a phenethyl
substituent (40) led to a considerable loss in potency, consistent with the limited size of the C2-pocket.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 46
IC50
1
2
3
4
5
6
7
8
Entry
R3
R2
A
B
HN
N
(nM)
HN
N
N
N
N
N
N
N
O
P
O
P
35
(R)-OH
48
2.2
N
R3
O
P
O
P
O
R3
O
O
HO
OH OH
O
HO
OH OH
R2
HO
R2
HO
36
37
38
39
40
(R)-OH
(S)-F
(S)-F
(S)-F
(S)-F
OMe
IC50
(nM)
IC50
(nM)
Entry
30
R3
R2
Entry
R3
R2
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)-F
4.0
1.3
10.0
1.5
41
(S)-F
(S)-F
(S)-F
(S)-F
0.6
1.4
3.3
202
OMe
Me
31
(R)-OH
32
33
34
(R)-OH
(S)-F
12.0
18.0
5.0
42
43
44
Me
OH
2.0
CF3
OH
OH
O
Me
(R)-OH
18.0
N
H
Table 2. A. Modification of C2 position (R3) with constant C6-benzyl group. B. Modification of C2 position (R3) with constant C6-
cyclopentyl group.
With a basic understanding of the nucleoside SAR and potent compounds in hand, we began optimization of PK
parameters in pursuit of a drug which was suitable for infrequent IV-administration. Compounds in the β-fluoroarabinose
series were generally characterized by favorable pharmacokinetic profiles, exhibiting very low clearance and long half-
lives in rat (Table 3). To further optimize for low clearance compounds, four potent analogs with varied C6-substituents
(R1, Table 3) were chosen. Compounds with benzylamine at R1 (16, Table 3) exhibited reduced clearance rates relative to
cyclopentylamine analogs (12). Clearance in rats could be further reduced by substituting the C6 position with (S)-(−)-α-
methylbenzylamine (20, CL = 0.021 L/h/kg). The stereochemistry at the benzylic position was critical for improving
metabolic stability. Interestingly, the opposite diastereomer, featuring a (R)-(+)-α-methylbenzylamine (18) at C6, displayed
higher clearance than the unsubstituted benzylamine analog (16).
To continue to explore opportunities to improve potency and PK properties, we began investigating alternative
heterocycles to replace the adenine base. In the β-fluoroarabinose series, replacement of adenine with a
pyrazolopyrimidine heterocycle provided compounds 45 - 48, all of which displayed equivalent or reduced potency against
CD73. Gratifyingly, this relationship of C6-substitutent and clearance was conserved relative to compounds featuring the
ACS Paragon Plus Environment

Page 9 of 46
Journal of Medicinal Chemistry
adenine base. This clear and predictable relationship was integral in our optimization towards very low clearance CD73
inhibitors.
1
2
3
4
5
6
R1
7
Me
Me
8
9
Nucleoside Core
Assay
HN
HN
HN
HN
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
R1
N
ID
12
16
20
18
N
N
hCD73 IC50
0.4
0.7
4.1
1.6
O
P
O
(nM)
N
Cl
O
P
HO
O
OH OH
CL (rat)
(L/h/kg)
0.15
45
0.060
46
0.021
47
0.28
48
HO
F
R1
N
ID
hCD73 IC50
N
N
N
0.15
5.9
5.4
2.1
O
P
O
P
(nM)
N
Cl
Cl
O
O
HO
HO
O
CL (rat)
(L/h/kg)
OH OH
0.27
49
nd
50
0.029
51
0.56
52
HO
HO
F
R1
N
ID
hCD73 IC50
N
0.027
0.052
0.045
0.058
O
P
O
P
(nM)
N
O
OH OH
CL (rat)
(L/h/kg)
0.63
0.20
0.064
0.17
OH
Table 3. Concise optimization of potency and pharmacokinetic properties by variation of C6 substituent (R1) and purine
heterocycle. Rat clearancedetermined in maleSD rats following an IVdoseof 0.5 mg/kg.
The analogous pyrazolopyrimidines (49 - 52) featuring a ribose core exhibited remarkably improved potency.
Relative to analogs possessing the β-fluoroarabinose core, ribose analogs were 10 to 100-fold more potent. For example,
compound 51 displayed extraordinary potency (IC50 = 45 pM) and very low clearance in rats (0.064 L/h/kg). With this
discovery, the development of a long-active IV-administered drug appeared within reach.
The introduction of the pyrazolopyrimidine core was, however, associated with worsened chemical stability.
Anomeric cleavage of the heterocycle was prevalent under acidic conditions.27 Furthermore, the electrophilicity at the C2
position was evidenced by displacement of the C2-Cl by water under basic conditions. Lastly, metabolic profiling identified
a potential for oxidation of the unsubstituted C6-benzylamine group. We believed we could mitigate each of these issues
by further modification of the purine heterocycle and modification of the C6-benzylamine moiety. We investigated three
heterocycles, pyrazolopyrimidine (control, Table 4A), pyrrolopyrimidine (Table 4B), and pyrazolopyridine (Table 4C) in
combination with ortho, meta or para-fluorine substituted (S)-(−)-α-methylbenzylamines at the C6 position. Incorporation
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 46
of fluorine at these positions had little effect on potency. The pyrazolopyridine analogs (55 and 60) demonstrated the
lowest in vivo rat clearance of the analogs examined. For example, 60 exhibited remarkably low clearance (0.009 L/h/kg)
in rats. Furthermore, 55 and 60 exhibited improved chemical stability under both acidic and basic conditions. Given their
favorable properties, 55 and 60, as well as 51, were selected for further characterization.
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
R1
R1
N
R1
N
A
B
C
O
P
O
P
X
O
N
N
HO
O
N
N
OH OH
N
N
N
Cl
N
Cl
Cl
HO
F
OH
X =
X =
X =
IC50
(nM)
CLrat
(L/h/kg)
IC50
(nM)
CLrat
IC50
(nM)
CLrat
(L/h/kg)
R1
ID
ID
ID
(L/h/kg)
Me
53
0.070
0.04
0.05
-
54
57
59
0.028
0.032
0.031
0.05
55
-
0.041
0.02
HN
Me
Me
F
56
58
0.031
0.095
0.04
0.03
-
-
HN
HN
60
0.115
0.009
F
Table 4. Further optimization of purine heterocycle and C6-substituent in parallel. Rat clearance
determined in maleSD rats following an IVdoseof 0.5 mg/kg.
CD73 is expressed on the surface of cancer cells as well as a variety of cells which comprise the immune infiltrate.
Additionally, soluble CD73 is abundant in human serum. As such, we examined the potency of 51, 55 and 60 against human
and mouse CD8+ T cells, human peripheral blood monocytic cells (PBMCs), and soluble recombinant CD73. All compounds
retained potency against soluble CD73 and were extraordinarily potent against human CD8+ T cells, with 51 and 55
exhibiting IC50 values of less than 10 pM. Similar potencies were observed for human PBMCs. In contrast, a near 100-fold
loss in potency was observed against mouse CD8+ T cells for the three compounds. The relative ranking of the three
compounds was conserved across all cell-based assays. In spite of the extraordinary potency, 55 was confirmed to be a
reversible, competitive inhibitor of CD73 with a Ki of 5 pM.28 None of the compounds showed any off-target activity against
the ecto-nucleotidases NTPDase-1 (CD39), -2, -3, and -8 up to 10 µM.29 Compounds 51, 55, and 60 were also screened
against 60 targets in a wide ligand profile screen (CEREP) with no off-target activity observed up to 10 µM.
ACS Paragon Plus Environment

Page 11 of 46
Journal of Medicinal Chemistry
hCD73-
Soluble
(IC50, nM)
hCD8+
T Cells
(IC50, nM)
mCD8+
T Cells
(IC50, nM)
1
2
3
hPBMC
(IC50, nM)
CD39
(IC50, µM)
NTPDase 2 NTPDase 3 NTPDase 8
Compound
(IC50, µM)
(IC50, µM)
(IC50, µM)
51
55
60
0.053
0.043
n.d.
0.0085
0.011
0.021
> 10
> 10
> 10
> 10
> 10
> 10
> 10
> 10
> 10
> 10
> 10
> 10
4
5
6
7
0.0027
0.0080
0.010
0.16
0.66
0.83
8
9
Table 5. Potency of select compounds against soluble CD73 and primary immune cell types. Selectivity against related
ecto-nucleotidases.
10
11
12
13
14
15
16
Mouse, rat, dog and monkey IV-administered PK data were collected for 51, 55, and 60 (Table 6). Low clearance
was observed across all species for each compound. Allometric scaling was employed to project human PK parameters.
The pyrazolopyridine analogs, 55 and 60, were projected to possess 98- and 167-hour human half-lives, respectively,
suitable for intravenous administration every 2 weeks (Q2W). Due to its superior potency in vitro and favorable
pharmacokinetic properties, 55 was selected for further characterization and ultimately became our clinical development
candidate AB680.
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
51
AB680 (55)
60
CL
Vss
T1/2
fu
CL
Vss
T1/2
fu
CL
Vss
T1/2
fu
(%)
Species
Mouse
Rat
(L/h/kg)
(L/kg)
(h)
(%)
(L/h/kg)
(L/kg)
(h)
(%)
(L/h/kg)
(L/kg)
(h)
2.0
1.7
13
28
27
3.5
5.3
22
27
98
6.5
9.6
15
2.9
0.076
0.064
0.18
0.12
0.84
0.096
0.14
0.025
0.020
0.12
0.89
0.17
8.3
0.015
0.0093
0.034
0.11
0.12
0.63
0.078
0.13
0.65
0.16
10.2
0.34
0.29
0.95
18.7
0.20
0.89
0.12
1.3
Dog
0.071
0.050
Monkey
34
0.0024
0.0036
0.0025
0.0012
0.099
0.17
0.59
0.42
0.0017
0.00054
Human
(projected)
167
Table 6. Pharmacokinetic data for select compounds in preclinical species following an IV dose of 0.5 mg/kg. Projected human PK
parameters determined by allometric scaling. Vss prediction determined by Øie-Tozer method.
AB680 was assessed for its potential to cause drug-drug interactions (Table 7). No inhibition was observed against
a panel of CYP isoforms up to 100 µM and no potential for time-dependent CYP inhibition was observed. Potential for CYP
induction, assessed via enzymatic activity, mRNA expression levels, or PXR reporter gene assay, was not observed.
Furthermore, no evidence for cardiovascular liabilities (hERG, IC50 > 10 µM AB680) or genotoxicity (AMES negative) was
found.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 46
1
2
3
4
5
6
7
Asp-085
His-220
Asp-036
Assay
Result
Phe-417
His-038
CYP inhibition
(1A2, 2C9, 2C19, 2D6, 3A4)
All > 100 µM
8
9
P450 TDI
(1A2, 2C9, 2C19, 2D6, 3A4)
No issues
No issues
IC50 > 10 µM
No issues
Phe-500
His-243
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
AMES
Asn-245
hERG
Asp-506
(automatic patch clamp)
Asn-390
Arg-354
CYP Induction, mRNA
(1A2, 2B6, 3A4)
Figure 4. X-ray structure of AB680 bound to CD73 (closed
Table 7. Additional in vitro profiling of AB680
form), PDB ID: 6Z9D.
Finally, X-ray co-crystallization of AB680 with CD73 revealed a binding mode analogous to AMPCP and 16,
described above. The density map indicated flexibility within the α-methylbenzylamine moiety, suggesting this group may
bind as a mixture of rotational isomers. The major conformation (63% abundance) is shown in Figure 4. The π-stacking
interaction of AB680 is slightly altered relative to that of AMPCP with the nucleobase of AB680 being shifted away from
F417 towards F500. As a result, F417 no longer forms a parallel face-to-face π-stacking interaction with the nucleobase of
AB680 but is tilted toward an edge-to-face interaction. As a result of this conformational change, F417 forms an additional
face-to-face π-interaction with the fluorphenyl substituent of AB680. These interactions may be energetically more
favorable and may partially explain the potency gain observed by incorporation of this nucleobase and the phenyl
substituent.30
Based on the extraordinary potency of AB680 in vitro and excellent pharmacokinetic profile in preclinical species,
AB680 was advanced to a phase I clinical study in healthy volunteers. This placebo-controlled study assessed the safety,
tolerability, and PK/PD profile of AB680. Here we present the initial pharmacokinetic data of AB680 in humans. AB680 was
well tolerated when administered as a single IV dose across the range of 0.1 to 25 mg (6 cohorts). All treatment-emergent
adverse events were mild or moderate in severity, with no clear pattern of toxicity at these dose levels. AB680 displayed
low clearance and a long half-life (67 – 74 hours) following an IV infusion of 16 or 25 mg over 30-60 minutes (Figure 6).
The human PK profile of AB680 is consistent with the intended Q2W dosing schedule and validates the design strategy
employed to discover a long-acting drug for IV-administration.
ACS Paragon Plus Environment

Page 13 of 46
Journal of Medicinal Chemistry
Phase I HV Study - Pharmacokinetics B.
A.
10000
1
2
Cohort
n
16 mg
6
25 mg
6
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
1000
100
10
Cmax (ng/mL)
CL (L/h)
Vss (L)
3,910
0.0672
6.38
5,170
0.0685
6.98
T1/2 (h)
66.6
74.2
25 mg (n = 6)
16 mg (n = 6)
1
0
48
96
144 192 240 288 336
Time (h)
Figure 6. A. Mean (SD) plasma concentrations of AB680 versus time in healthy volunteers (HV)
following an IV-infusion over 30 minutes at dose levels of 16 and 25 mg/kg. B. Non-compartmental PK
parameters following a singleIVdose of AB680.
Chemistry. Synthesis of representative examples 16 and AB680. The synthesis of analogs possessing a β-fluoroarabinose
or ribose core followed a similar general strategy; building the nucleotide core followed by phosphonylation as the final
step.31 In the case of β-fluoroarabinose analogs, such as 16 (Scheme 1), glycosylation of the 2,6-dichloro-
imidazolopyrimidine was achieved by treatment with Cs2CO3 followed by addition of benzoate protected 2-fluoro-α-D-
arabinofuranosyl bromide (61) to the reaction mixture.32 This cleanly provided 62 with complete regio- and
diastereomeric control. SNAr displacement of the C6-chloride was achieved by treatment with benzylamine and Et3N in
EtOH at elevated temperature. No displacement of the C2-chloride was observed. Global benzoate deprotection was
realized by treatment with K2CO3 in MeOH at room temperature to provide the penultimate nucleoside 63 in 70 % yield.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 46
Cl
1) HMDS, (NH4)2SO4
reflux
Cl
1
2
3
4
5
6
7
8
Cl
N
Cl
2)TMSOTf, MeCN, r.t.
N
N
N
N
N
N
Cl
N
N
O
O
AcO
OAc
N
N
Cl
H
AcO
N
H
O
Cl
Br
N
(29%)
N
Cl
BzO
O
64
65
AcO
Me
AcO
Me
OAc
OAc
BzO
Cs2CO3, ACN
r.t.
BzO
F
F
1)
F
BzO
F
61
62
(77%)
Et3N, EtOH
110 °C;
HN
H2N
66
HN
NH3, MeOH, rt
9
1) BnNH2, Et3N
EtOH, 70 °C
(70%, two-
steps)
N
N
N
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
N
Cl
O
2)
OMe
MeO
N
HO
N
Cl
O
2) K2CO3, MeOH
r.t
HO
67
63
Me
F
p-TsOH, acetone
(29%)
O
O
HN
HO
F
HN
N
N
O
O
1)
Cl
O
P
O
N
O
O
P
N
P
P
P
O
Cl
Cl
O
Cl
N
N
Cl
O
N
P
Cl
Cl
Cl
Cl
Cl
O
P
P
HO
O
O
HO
OH OH
16
OH
OH
(MeO)3PO, 0 °C
(22%)
(MeO)3PO, 0 °C
AB680
HO
F
HO
OH
2)
H2O
-20 to 0 °C
Scheme 1. Preparation of 16 - Representative synthesis of
-fluoroarabinoseanalogs.
(50%)
Scheme 2. Synthesis of AB680.
5’-phosphorylation and phosphonylation of nucleosides has been reported to be rapid and regioselective in the presence
of PO(OMe)3 without the necessity of protecting secondary alcohols.33 Indeed, treatment of 63 with excess
methylenebis(phosphonic dichloride) in PO(OMe)3 resulted in the selective phosphonylation of the 5’-hydroxyl group.
Following quenching the phosphonyl chloride with aqueous buffer, 16 was isolated by reverse phase preparative HPLC.
The synthesis of AB680 and related ribose analogs employed a similar strategy. The pyrazolopyridine nucleobase
presented unique challenges in the glycosylation step. Selectivity for pyrazole-N1 glycosylation was achieved by exposure
of 64 to catalytic ammonium sulfate in refluxing HMDS following established methods. After removal of excess HMDS, the
crude silylated heterocycle was reconstituted in MeCN and alkylated with β-D-ribofuranose-1,2,3,5-tetraacetate
promoted by TMS-triflate. Careful separation of the minor undesired anomer was required by column chromatography.
Displacement of the C6-chloride with benzylamine 66 was achieved by heating in EtOH with excess Et3N at 110 °C.
Subsequent removal of the 2’,3’,5’-acetate groups was rapid in the presence of methanolic ammonia. Following acetonide
protection of the 2’,3’-hydroxyl groups, phosphonylation of the 5’-position was achieved under conditions analogous to
those described above to prepare 16. Acetonide deprotection, promoted by acid generated by quenching the intermediate
phosphonyl chloride with water, afforded AB680. The remaining pyrazolopyridine-containing analog 60 was prepared via
an analogous route.
ACS Paragon Plus Environment

Page 15 of 46
Journal of Medicinal Chemistry
Synthesis of C6 modified analogs (Table 1). C6-modified analogs were readily prepared in a high throughput manner via
nucleophilic aromatic substitution with amines or alcohols (Scheme 3). Within the β-fluoroarabinose scaffold, benzoyl
protected 2,6-dichloroadenosine intermediate 62 was treated with a variety of primary and secondary amines and
triethylamine in EtOH at 70 °C. In every instance SnAr occurred exclusively at the C6 position in moderate to good yield.
Removal of the 3’,5’-benzoyl protecting groups was achieved by treatment with K2CO3 in MeOH providing variably C6-
amine substituted nucleosides 68. The corresponding nucleotides were prepared under conditions identical to those
employed to prepare 16 (Scheme 1). Ribose nucleosides 71 were prepared similarly from commercially available dichloride
70 (Scheme 3).
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
R1
Cl
N
HN
N
O
P
O
P
1) R1-NH2, Et3N
EtOH, 70 °C
N
N
N
N
N
N
Cl
Cl
Cl
Cl
3, 6, 7, 8, 10, 12, 14,
16, 18, 20, 22-28
Cl
Cl
Cl
O
O
BzO
HO
HO
2) K2CO3, MeOH
r.t
(MeO)3PO, 0 °C
BzO
F
HO
HO
F
Bn
N
68
62
O
N
O
P
O
P
N
N
HO
Cl
Cl
Cl
Cl
O
29
NaH, r.t.
(97%)
(MeO)3PO, 0 °C
F
69
R1
N
Cl
N
HN
N
O
P
O
P
N
N
N
N
N
Cl
Cl
R1-NH2, Et3N
EtOH, 60 °C
Cl
Cl
Cl
Cl
O
O
4, 9, 11, 13, 15,
17, 21
HO
HO
(MeO)3PO, 0 °C
HO
OH
HO
OH
70
71
Scheme 3. General synthesis of C6-modified nucleotide analogs. See experimental section for details.
Synthesis of C2 modified analogs (Table 2). Synthetic access to C2 modified analogs was generally accomplished by
nucleophilic substitution of a suitable aryl halide or transition metal-mediated catalysis. Compounds 34-36 were prepared
by the former mechanism. Treatment of nucleoside 72 with excess methylamine in water at 60 °C provided the desired
C2-substituted nucleoside in 85% yield (Scheme 4). The analogous C2-pyrrolidine analog (precursor to 35) was prepared
under similar conditions. Sodium methoxide was also a competent nucleophile. Treatment with excess NaOMe in MeOH
displaced the C2-chloride in good yield providing nucleoside 74. Phosphonylation under standard conditions provided the
final nucleotide analogs 34 and 36.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 46
1
2
3
4
5
6
7
8
9
HN
N
HN
N
O
P
O
P
N
N
Cl
Cl
N
N
40% MeNH2, H2O
Cl
Cl
Me
34
N
Cl
N
N
O
O
H
(MeO)3PO, 0 °C
(15%)
60 °C (85%)
HO
HO
HO
HO
OH
OH
72
73
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
HN
N
HN
N
O
P
O
P
N
N
N
N
Cl
Cl
N
N
25% NaOMe, MeOH
60 °C (73%)
Cl
Cl
36
Cl
OMe
O
O
(MeO)3PO, 0 °C
(14%)
HO
HO
HO
HO
OH
OH
72
74
Scheme 4. Synthesis of C2-modified analogs 34 and 36 via SnAr with amine and alcohol nucleophiles
The remaining analogs presented in Table 2 were accessed via palladium-catalyzed cross coupling of suitably
functionalized 2-halo-adenine intermediates and a nucleophilic coupling partner. The C2-methyl analog 31 was prepared
from the corresponding C2-iodide 75 by Stille reaction with Me4Sn and Pd(PPh3)4 in NMP at 120°C followed by
deprotection and phosphonylation. C2-Ethyl analog 32 was prepared analogously via Suzuki coupling of vinylboronic acid
pinacol ester with 72 followed by phosphonylation and reduction.
HN
N
HN
N
O
P
O
P
N
N
N
N
1)
2)
Me4Sn, Pd(PPh3)4
NMP, 120 °C
Cl
Cl
N
N
Cl
Cl
31
I
CH3
O
O
K2CO3, MeOH, r.t
(60%)
AcO
(MeO)3PO, 0 °C
HO
AcO
OAc
HO
OH
75
76
HN
N
HN
N
O
P
O
P
N
N
N
N
1)
BPin
N
N
Cl
Cl
Pd(PPh3)4, K2CO3
Cl
Cl
Cl
O
O
32
HO
HO
DME:H2O (9:1)
85 °C (72%)
(MeO)3PO, 0 °C (8%)
2)
HO
Pd/C, H2, MeOH
OH
HO
OH
(35%)
72
77
Scheme 5. Synthesis of C2-modified analogs 31 and 32.
The syntheses of β-fluoroarabinose C2-modified nucleosides 79-82 are described in Table 7. Each was prepared
from compound 78 via Suzuki or Sonogashira coupling under standard conditions followed by benzoate deprotection with
K2CO3 in MeOH. Subsequent alkyne hydrogenation was required to prepare C2-phenethyl nucleoside analog 82.
Nucleosides 79-82 were converted to their corresponding nucleotide analogs 37-40 under standard conditions.
ACS Paragon Plus Environment

Page 17 of 46
Journal of Medicinal Chemistry
HN
HN
1
2
3
4
5
6
N
N
N
N
N
N
78
Cl
R2
O
N
Table 7
N
O
BzO
HO
79-82
BzO
F
HO
F
7
8
9
Pdt #
R2
Conditions
Phenylboronic acid,
Pd(PPh3)4, K2CO3, THF, H2O
Yield (%)
1)
8
79
80
81
(two steps)
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
2)
K2CO3, MeOH
1) K-benzyltrifluoroborate,
PdCl2(dppf), Cs2CO3, THF,H2O
40
Ph
2)
K2CO3, MeOH
86
43
1) TMS-acetylene,Pd(PPh3)2Cl2,
CuI, Et3N, DMF
2)
1)
K2CO3, MeOH
42
92
Phenylacetylene,
Pd(PPh3)2Cl2, CuI, Et3N, DMF
82
2)
3)
K2CO3, MeOH
Pd/C, H2
Ph
79
(steps 2-3)
Table 7. Summary of conditions employed for C2-modification
of β-fluoroarabinose nucleosides. Products 79-82 were
converted to nucleotides 37-40, respectively, under standard
conditions.
The analogs described in Table 2B were prepared in a divergent manner from C2-aldehyde 84 which was prepared
via Suzuki coupling of 83 with trans-2-phenylvinylboronic acid and subsequent Lemieux-Johnson oxidative olefin cleavage.
Intermediate 84 was reduced with NaBH(OAc)3 to give benzyl alcohol 85. Activation with p-TsCl followed by methanolysis
provided 86, the nucleoside precursor to 41.
B(OH)2
HN
N
1)
2)
HN
N
HN
N
Ph
N
N
Pd(PPh3)4, Na2CO3
THF, H2O (72%)
N
N
N
N
N
N
N
NaBH(OAc)3
DCE
OH
Cl
CHO
HN
O
O
O
K2OsO4•2H2O
NaIO4, 2,6-lutidine
THF, H2O (97%)
BzO
BzO
BzO
BzO
BzO
F
F
BzO
F
83
84
85
O
O
P
1) p-TsCl, Et3N
P
Cl
Cl
N
N
CH2Cl2
Cl
Cl
OMe
41
N
N
O
(MeO)3PO, 0 °C
2) K2CO3, MeOH
HO
(22%, 3-steps)
86
HO
F
Scheme 6. Synthesis of 41.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 46
Aldehyde 84 was also treated independently with cyclopropyl, phenyl, and tetrahydropyranyl Grignard reagents to provide
secondary alcohols 87-89, respectively, as a mixture of diastereomers. Excess organomagnesium reagent was utilized to
remove the 3’ and 4’-benzoate groups in situ. Phosphonylation under standard conditions provided analogs 42-44.
1
2
3
4
5
6
7
8
9
HN
N
HN
N
O
P
O
P
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
N
N
N
N
N
Cl
Cl
RMgBr
Cl
Cl
OH
42 - 44
CHO
O
O
HO
BzO
(MeO)3PO, 0 °C
THF, -78 °C
R
HO
F
BzO
F
84
(87)
(88)
(89)
R =
R =
R =
O
Scheme 7. Synthesis of 42-44.
Representative syntheses of analogs with pyrazolopyrimidine and pyrrolopyrimidine nucleobases (Tables 3 and 4). The
syntheses of ribose and β-fluoroarabinose containing a pyrazolopyrimidine nucleobase began from commercially available
2,6-dichloropyrazolopyrimidine 90 (Scheme 8). In contrast to β-fluoroarabinose adenine analogs, such as 16, nucleophilic
substitution of the 6-position with alkyl groups was performed prior to glycosidic bond formation. This was necessary due
to the high relative electrophilicity of 90. In fact, exposure of 90 to Cs2CO3 caused rapid polymerization of 90 at room
temperature.
In one example, treatment of 90 with cyclopentylamine rapidly formed 91. Glycosylation of 91 was achieved by
treatment with 2-fluoro-α-D-arabinofuranosyl bromide 61 in the presence of Cs2CO3. Under these conditions, both
pyrazolopyrimidine N1- and N2-alkylated products were generated, which were readily separated after deprotection with
K2CO3 in MeOH. Nucleoside 92 was converted to 45 under phosphonylation conditions identical to those described for
the synthesis of 16. Analogs 46-48 were prepared by an analogous route where cyclopropylamine was replaced with the
suitable benzylamine in the first step of the sequence.
ACS Paragon Plus Environment

Page 19 of 46
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
O
Br
61
HN
N
BzO
Cl
N
HN
N
N
1)
N
H2N
BzO
F
N
N
N
N
N
N
Cl
Cl
O
N
Et3N, MeCN
rt
N
Cs2CO3, MeCN
rt
K2CO3, MeOH
HO
HO
N
Cl
Cl
N
H
H
90
91
2)
HO
HO
F
92
(41%, three-steps)
HN
N
Cl
N
HMDS, (NH4)2SO4
reflux;
9
Et3N, MeOH
N
N
10
11
12
13
14
15
16
17
18
19
20
21
N
O
Cl
O
OAc
O
AcO
AcO
H2N
0 °C to rt;
NH3, MeOH
AcO
OAc
OH
AcO
OAc
94
93
TMS-OTf, CH3CN, r.t. (82%)
Scheme 8. Synthesis of pyrazolopyrimidinenucleosides.
Pyrazolopyrimidine nucleosides featuring a ribose moiety as exemplified by 94 were accessed via a synthetic
sequence analogous to that employed for the synthesis of AB680 (vide supra). In short, 90 was refluxed with HMDS in the
presence of catalytic ammonium sulfate followed by treatment with β-D-ribofuranose-1,2,3,5-tetraacetate and TMS-
triflate to prepare protected nucleoside 93. Treatment with cyclopentylamine and Et3N in MeOH resulted in rapid and
selective displacement of the C6-chloride. Global deprotection of the intermediate 2’,3’,5’-triacetate was achieved by
exposure to ammonia in MeOH, affording nucleoside 94. Under standard phosphonylation conditions described above,
94 was converted to 49. Analogs 50-52, 53, 56, and 58 were prepared by an analogous route where cyclopropylamine was
replaced with the suitable benzylamine.
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
Me
F
Me
F
O
Cl
N
HN
N
Br
AcO
H2N
Cl
N
N
N
AcO
OAc
N
N
N
Cl
Cl
O
O
Cs2CO3, MeCN, rt
(39%)
Et3N, EtOH
80 °C;
AcO
HO
N
Cl
H
NH3, MeOH
95
AcO
OAc
HO
OH
97
96
Scheme 9. Synthesis of pyrrolopyrimidinenucleosides
Analogs containing a pyrrolopyrimidine nucleobase (54, 57, and 59) were prepared as described in Scheme 9. 2,6-
Dichloropyrrolopyrimidine was glycosylated by treatment with D-arabinofuranosyl bromide-2,3,5-triacetate and Cs2CO3
in MeCN at room temperature affording 96 in moderate yield. Nucleophilic substitution of the C6-chloride with (S)-1-(2-
fluorophenyl)ethanamine was achieved by heating in EtOH in the presence of Et3N. Subsequent treatment with ammonia
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 46
in MeOH deprotected the intermediate 2’,3’,5’-triacetate to afford nucleoside 97. Under standard phosphonylation
conditions, described above, 97 was converted to 54. Analogs 57 and 59 were prepared by an analogous route.
1
2
3
4
5
6
7
8
Conclusion. Inhibition of CD73 eliminates a major pathway of ADO production in the TME and can reverse ADO-mediated
immune suppression. A series of potent and selective CD73 inhibitors were optimized using structure-based design and
interrogation of structure activity relationships. Trend analysis of rat IV pharmacokinetic data established clear and
consistent structure-pharmacokinetic relationships. Taken together, structure-activity and pharmacokinetic relationships
led to the discovery of AB680, which demonstrates extraordinary potency and very low clearance across preclinical
species. AB680 is currently being evaluated in Phase 1 clinical trials. Initial data show AB680 is well tolerated and exhibits
a pharmacokinetic profile suitable for Q2W IV-administration in human. Even though AB680 was optimized for parenteral
delivery, an extensive formulation effort identified proprietary formulations that allow for the oral delivery of AB68034.
Details of this oral formulation will be published in due course.
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
Experimental.
General Chemistry. All reactions were performed using a Teflon-coated magnetic stir bar at the indicated temperature
and were conducted under an inert atmosphere when stated. All chemicals were used as received. Reactions were
monitored by TLC (silica gel 60 with fluorescence F254, visualized with a short wave/long wave UV lamp) and/or LCMS
(Agilent 1100 series LCMS with UV detection at 254 nm using a binary solvent system [0.1% TFA in MeCN/0.1% TFA in H2O]
using either of the following column: Agilent Eclipse Plus C18 [3.5 µm, 4.6 mm i.d. x 100 mm]). Flash chromatography was
conducted on silica gel using an automated system (CombiFlash RF+ manufactured by Teledyne ISCO), with detection
wavelengths of 254 and 280 nm. Reverse phase preparative HPLC was conducted on an Agilent 1260 Infinity series HPLC.
Samples were eluted using a binary solvent system (0.1% TFA in MeCN/0.1% TFA in H2O) with gradient elution on a Gemini
C18 110 Å column (21.2 mm i.d. x 250 mm) with detection at 254 nm. Final compounds obtained through preparative
HPLC were concentrated through lyophilization. All reported yields are isolated yields. All assayed compounds were
purified to ≥95% purity as determined by HPLC or LCMS (Agilent 1100 series LCMS with UV detection at 254 nm using a
binary solvent system [0.1% TFA in MeCN/0.1% TFA in H2O or MeCN/50 mM sodium phosphate in water, pH 8.5] using
the following column: Agilent Eclipse Plus C18 column [3.5 µm, 4.6 mm i.d. x 100 mm]). 1H NMR spectra were recorded
on a Varian 400 MHz NMR spectrometer equipped with an Oxford AS400 magnet. Chemical shifts (δ) are reported as parts
ACS Paragon Plus Environment

Page 21 of 46
Journal of Medicinal Chemistry
per million (ppm) relative to residual undeuterated solvent as an internal reference. The abbreviations s, br. s, d, t, q, dd,
dt, ddd, and m stand for singlet, broad singlet, doublet, triplet, quartet, doublet of doublets, doublet of triplets, doublet
of doublet of doublets, and multiplet, respectively.
1
2
3
4
5
6
7
8
Clinical Study.
9
The phase I clinical study was a first-in-human, double-blind, randomized, placebo-controlled combined single-ascending-
dose (SAD) and multiple-ascending-dose (MAD) study to evaluate the safety, tolerability, PK and potential PD effects of
AB680 in healthy volunteers. Participants were randomly to receive AB680 (n=6) or matching placebo (n = 2) in each of 7
dosing cohorts in the SAD part and a single dose cohort in the MAD part. In the SAD part of the study, participants received
a single IV infusion of 0.1, 0.6, 2, 4, 8, 16, 25 mg AB680 or placebo. In the MAD part of the study, participants received IV
infusion of 8 mg AB680 or placebo once daily on 3 days (Days 1, 8, and 15).
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
Compound Synthesis.
[(2R,3R,4S,5R)-3-Benzoyloxy-5-(2,6-dichloro-9H-purin-9-yl)-4-fluorotetrahydrofur-2-yl]methyl benzoate (62). 2,6-
dichloropurine (3.6 g, 18.8 mmol) was dissolved in 90 mL of MeCN and treated with Cs2CO3 (7.5 g, 23 mmol, 1.2 equiv.).
The mixture was stirred at room temperature for 30 min.
(2R,3R,4S,5R)-2-(Benzoyloxymethyl)-5-bromo-4-
fluorotetrahydrofur-3-yl benzoate (8.75 g, 21 mmol, 1.1 equiv.) was dissolved in 100 mL of MeCN and added to the mixture
dropwise via an addition funnel. The mixture was allowed to stir overnight at room temperature. The mixture was filtered
on a pad of silica gel and concentrated. The residue was adsorbed on silica and purified using column chromatography
1
(hexanes / EtOAc) to provide the product (62) as a white solid in 77% yield (7.72g). H NMR (400 MHz, Chloroform-d) δ
8.39 (d, J = 3.0 Hz, 1H), 8.10 (ddt, J = 8.5, 3.1, 0.9 Hz, 4H), 7.74 – 7.36 (m, 6H), 6.64 (dd, J = 21.8, 2.8 Hz, 1H), 5.83 – 5.69
(m, 1H), 5.40 (ddd, J = 49.9, 2.8, 0.8 Hz, 1H), 4.89 – 4.77 (m, 2H), 4.62 (q, J = 4.0 Hz, 1H). ESI MS [M+H]+ for C24H17Cl2FN4O5,
calcd 531.1, found 531.1.
(2R,3R,4S,5R)-5-(9a-benzyl-2-chloro-9-adenineyl)-4-fluoro-2-(hydroxymethyl)tetrahydrofuran-3-ol (63). Compound 62
(9.0 g, 17 mmol), benzylamine (3 mL, 26 mmol, 1.5 equiv.), and Et3N (5 mL, 34 mmol, 2.0 equiv.) in anhydrous EtOH (60
mL) was stirred at 70 °C for 4 hours. The reaction mixture was then cooled to room temperature and the product was
collected by filtration and used without further purification (white solid, 8.9 g, 87%). ESI MS [M+H]+ for C31H25ClFN5O5,
calcd 602.2, found 602.0.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 46
The above product (10.2 g, 17 mmol) and K2CO3 (7 g, 51 mmol, 3 equiv) were dissolved in 170 mL of MeOH and stirred at
room temperature for 4 hours. The reaction mixture was then filtered and concentrated on a pad of silica gel. The reaction
mixture was purified using column chromatography (CH2Cl2 / MeOH) to provide the product 63 as a white solid in 80%
1
2
3
4
5
6
7
8
1
yield (5.3 g): H NMR (400 MHz, DMSO-d6) δ 8.97 (t, J = 6.3 Hz, 1H), 8.31 (d, J = 2.0 Hz, 1H), 7.36 – 7.18 (m, 5H), 6.34 (dd,
9
J = 13.6, 4.7 Hz, 1H), 5.23 (dt, J = 52.6, 4.3 Hz, 1H), 4.66 (q, J = 7.3, 5.7 Hz, 2H), 4.43 (dt, J = 19.0, 4.8 Hz, 1H), 3.84 (q, J = 4.9
Hz, 1H), 3.65 (tq, J = 12.0, 6.2, 5.2 Hz, 2H). ). ESI MS [M+H]+ for C17H18ClFN5O3, calcd 394.1, found 394.1.
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
[[(2R,3R,4S,5R)-5-[6-(benzylamino)-2-chloropurin-9-yl]-4-fluoro-3-hydroxyoxolan-2-yl]methoxy-
hydroxyphosphoryl]methylphosphonic acid (16). Compound 63 (800 mg, 2 mmol) was dissolved in trimethyl phosphate
(15 mL) and cooled to 0 °C (ice bath), then a cold solution of methylenebis(phosphonic dichloride) (2.5 g, 10 mmol, 5
equiv.) in trimethyl phosphate (5 mL) was added dropwise. The reaction mixture was stirred at 0 °C for 3 h, and was then
carefully quenched with 0.5 M triethylammonium bicarbonate solution (15 mL) and stirred at 0 °C for 15 min, and then 2
h at room temperature. The reaction mixture was purified by reverse phase HPLC (C18 column, 0 to 40% gradient of MeCN
1
and water with 0.1% TFA) to give the product 16 as a white solid in 22% yield (290 mg): H NMR (400 MHz, DMSO-d6) δ
8.99 (t, J = 6.3 Hz, 1H), 8.30 (d, J = 2.2 Hz, 1H), 7.40 – 7.18 (m, 5H), 6.38 (dd, J = 14.3, 4.6 Hz, 1H), 5.45 – 5.04 (m, 1H), 4.65
(t, J = 5.5 Hz, 2H), 4.54 – 4.42 (m, 1H), 4.19 (t, J = 6.1 Hz, 2H), 4.04 (t, J = 5.1 Hz, 1H), 2.26 (t, J = 20.5 Hz, 2H). ESI MS [M-
H]- for C18H21ClFN5O8P2, calcd 550.8, found 550.2.
[(2R,3R,4R,5R)-3,4-diacetoxy-5-(4,6-dichloro-1,2,7-triaza-1H-inden-1-yl)tetrahydrofur-2-yl]methyl acetate (65). 4,6-
Dichloro-1H-pyrazolo[3,4-d]pyrimidine 64 (7.0 g, 37 mmol) and ammonium sulfate (100 mg, 5.3 mmol, 1 mol %) were
dissolved in hexamethyldisilizane (40 mL). The mixture was heated to reflux and stirred for 3 h. Following this time, the
mixture was concentrated in vacuo and then dried under high vacuum (~1 h). The residue was then taken up in MeCN (70
mL), β-D-ribofuranose 1,2,3,5-tetraacetate (12.9 g, 40.7 mmol, 1.1 equiv) was added. This mixture was cooled to 0 °C and
TMSOTf (10 mL, 55.5 mmol, 1.5 equiv) was added dropwise. The reaction was stirred for an additional hour at 0 °C then
quenched by slow addition of saturated aqueous NaHCO3 (1 L) at 0 °C. The mixture was diluted with EtOAc. The aqueous
phase was extracted with EtOAc and then the combined organic extracts were washed with brine, dried over MgSO4 and
concentrated in vacuo. The residue was purified by column chromatography (SiO2, 0→40% EtOAc/Hex) to afford the
product 4 as a foam (4.8 g, 29%). 1H NMR (400 MHz, DMSO-d6) δ 8.56 (s, 1H), 7.72 (s, 1H), 6.49 (s, 1H), 5.99 – 5.75 (m, 1H),
ACS Paragon Plus Environment

Page 23 of 46
Journal of Medicinal Chemistry
5.73 – 5.54 (m, 1H), 4.48 – 4.25 (m, 3H), 4.12 – 3.96 (m, 1H), 2.10 (s, 3H), 2.07 (s, 3H), 1.97 (s, 3H). ESI MS [M+H]+ for
C17H17Cl2N3O7, calcd 446.0, found 446.2.
1
2
3
4
5
6
7
8
[(1R,5R,6R,8R)-8-{4-[(S)-1-(o-Fluorophenyl)ethylamino]-6-chloro-1,2,7-triaza-1H-inden-1-yl}-3,3-dimethyl-2.4.7-
trioxabicyclo[3.3.0]oct-6-yl]methanol (67). A solution of 65 (3.1 g, 6.9 mmol), (1S)-1-(2-fluorophenyl)ethylamine (1 g, 7.2
mmol) and triethylamine (1 mL, 7.2 mmol) in anhydrous EtOH (7 mL) was sealed in a screw-top flask and heated to 110 °C
overnight. After cooling to room temperature, 7M NH3/MeOH (5 mL) was added and the mixture was stirred overnight.
The reaction mixture was concentrated to dryness under reduced pressure, then reconstituted in acetone (10 mL). 2,2-
Dimethoxypropane (10 mL) and p-TsOH (1.6 g, 8.6 mmol) was added and the mixture was stirred overnight at room
temperature. The reaction mixture was concentrated to dryness under reduced pressure and purified by column
chromatography (SiO2, 0→50% EtOAc/Hex) to provide 67 (950 mg, 29% yield) as a white solid. 1H NMR (400 MHz, DMSO-
d6) δ 8.37 (br. s, 1H), 8.24 (d, J = 7.2 Hz, 1H), 7.37 (td, J = 7.8, 1.8 Hz, 1H), 7.32 – 7.25 (m, 1H), 7.23 – 7.11 (m, 2H), 6.21 (d,
J = 1.8 Hz, 1H), 5.98 (br. s, 1H), 5.26 (dd, J = 6.1, 1.9 Hz, 1H), 5.03 (br .s, 1H), 4.89 (dd, J = 6.2, 2.2 Hz, 1H), 4.85 (t, J = 5.8
Hz, 1H), 4.07 (ddd, J = 7.8, 5.9, 2.2 Hz, 1H), 3.52 – 3.37 (m, 1H), 3.33 – 3.22 (m, 1H), 2.47 (p, J = 1.8 Hz, 2H), 1.54 (d, J = 6.7
Hz, 3H), 1.51 – 1.44 (m, 3H), 1.33 – 1.27 (m, 3H). ESI MS [M+H]+ for C19H20ClFN4O4, calcd 423.1, found 423.3.
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
[[(2R,3S,4R,5R)-5-[6-chloro-4-[[(1S)-1-(2-fluorophenyl)ethyl]amino]pyrazolo[3,4-b]pyridin-1-yl]-3,4-dihydroxyoxolan-
2-yl]methoxy-hydroxyphosphoryl]methylphosphonic acid, AB680 (55). Compound 67 (200 mg, 0.43 mmol) was dissolved
in trimethyl phosphate (2 mL) and cooled to 0 °C (ice bath), then a cold solution of methylenebis(phosphonic dichloride)
(325 mg, 1.3 mmol, 3 equiv.) in trimethyl phosphate (1 mL) was added dropwise. The reaction mixture was stirred at 0 °C
for 3 h, and was then treated with water (1 mL) and stirred overnight at room temperature. The reaction mixture was
purified by reverse phase HPLC (C18 column, 0 to 40% gradient of MeCN and water with 0.1% TFA) to give the product
AB680 as a white powder in 50% yield (150 mg). 1H NMR (400 MHz, DMSO-d6) δ 8.38 (s, 1H), 8.23 (d, J = 6.9 Hz, 1H), 7.42
– 7.34 (m, 1H), 7.33 – 7.09 (m, 3H), 6.06 (d, J = 4.3 Hz, 1H), 5.97 (s, 1H), 5.04 (s, 1H), 4.53 – 4.47 (m, 1H), 4.25 (t, J = 4.7 Hz,
1H), 4.13 – 3.97 (m, 2H), 3.92 – 3.82 (m, 1H), 2.16 (d, J = 20.5 Hz, 2H), 1.56 (d, J = 6.4 Hz, 3H). ESI MS [M+H]+ for
C20H25ClFN4O9P2, calcd 581.1, found 581.2.
[[(2R,3R,4S,5R)-5-(6-amino-2-chloropurin-9-yl)-4-fluoro-3-hydroxyoxolan-2-yl]methoxy-
hydroxyphosphoryl]methylphosphonic acid (4). The title compound was synthesized in similar fashion to Example 16
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 46
using commercially available alcohol. 1H NMR (400 MHz, DMSO-d6) δ 8.28 (d, J = 2.2 Hz, 1H), 7.92 (s, 2H), 6.36 (dd, J =
14.3, 4.6 Hz, 1H), 5.26 (dt, J = 52.5, 4.3 Hz, 1H), 4.51 (dt, J = 18.6, 4.7 Hz, 1H), 4.19 (t, J = 6.0 Hz, 2H), 4.04 (t, J = 5.0 Hz, 1H),
2.26 (t, J = 20.5 Hz, 2H); MS: (ES) m/z calculated for C11H15ClFN5O8P2 [M-H]- 460.1, found 460.1.
1
2
3
4
5
6
7
8
[[(2R,3R,4R,5R)-5-(6-aminopurin-9-yl)-4-fluoro-3-hydroxyoxolan-2-yl]methoxy-hydroxyphosphoryl]methylphosphonic
9
1
acid (5): The title compound was synthesized in similar fashion to step c of Example 16 using corresponding alcohol: 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
NMR (400 MHz, DMSO-d6) δ 8.45 (s, 1H), 8.26 (s, 1H), 7.92 (s, 2H), 6.27 (dd, J = 17.2, 2.8 Hz, 1H), 5.50 (ddd, J = 52.5, 4.5,
2.8 Hz, 1H), 4.64 – 4.52 (m, 1H), 4.29 – 4.08 (m, 3H), 2.25 (t, J = 20.4 Hz, 2H). ESI MS [M+H]+ for C11H16FN5O8P2, calcd 428.1,
found 428.1.
[[(2R,3R,4S,5R)-5-[2-chloro-6-(methylamino)purin-9-yl]-4-fluoro-3-hydroxyoxolan-2-yl]methoxy-
hydroxyphosphoryl]methylphosphonic acid (6): The title compound was synthesized in similar fashion to Example 16: 1H
NMR (400 MHz, DMSO-d6) δ 8.36 (q, J = 4.6 Hz, 1H), 8.27 (s, 1H), 6.45 (brs, 2H), 6.37 (dd, J = 14.3, 4.6 Hz, 1H), 5.25 (dt, J =
52.4, 4.3 Hz, 1H), 4.50 (dt, J = 18.6, 4.6 Hz, 1H), 4.19 (t, J = 5.9 Hz, 2H), 4.04 (q, J = 5.2 Hz, 1H), 3.33 (brs, 1H), 2.93 (d, J =
4.5 Hz, 3H), 2.26 (t, J = 20.4 Hz, 2H). ESI MS [M-H]- for C12H17ClFN5O8P2, calcd 474.7, found 474.1.
[[(2R,3R,4S,5R)-5-[2-chloro-6-(ethylamino)purin-9-yl]-4-fluoro-3-hydroxyoxolan-2-yl]methoxy-
hydroxyphosphoryl]methylphosphonic acid (7): The title compound was synthesized in similar fashion to Example 16: 1H
NMR (400 MHz, DMSO-d6) δ 8.43 (t, J = 5.7 Hz, 1H), 8.26 (s, 1H), 7.28 (brs, 2H), 6.37 (dd, J = 14.3, 4.6 Hz, 1H), 5.25 (dt, J =
52.4, 4.3 Hz, 1H), 4.50 (dt, J = 18.5, 4.6 Hz, 1H), 4.19 (t, J = 6.1 Hz, 2H), 4.03 (q, J = 5.1 Hz, 1H), 3.87 (brs, 1H), 3.45 (m, 1H),
2.27 (t, J = 20.5 Hz, 2H), 1.17 (t, J = 7.2 Hz, 3H). ESI MS [M+H]+ for C13H19ClFN5O8P2, calcd 490.7, found 490.1.
[[(2R,3R,4S,5R)-5-[2-chloro-6-(cyclopropylamino)purin-9-yl]-4-fluoro-3-hydroxyoxolan-2-yl]methoxy-
hydroxyphosphoryl]methylphosphonic acid (8): The title compound was synthesized in similar fashion to Example 16: 1H
NMR (400 MHz, DMSO-d6) δ 8.59 (s, 1H), 8.28 (d, J = 2.1 Hz, 1H), 6.38 (dd, J = 14.2, 4.6 Hz, 1H), 5.26 (ddd, J = 52.5, 4.3, 4.3
Hz, 1H), 4.51 (dt, J = 18.5, 4.5 Hz, 1H), 4.19 (t, J = 6.1 Hz, 2H), 4.03 (q, J = 5.0 Hz, 1H), 2.98 (s, 1H), 2.36 – 2.15 (m, 2H), 0.82
– 0.48 (m, 4H). ESI MS [M−H]− for C14H18ClFN5O8P2, calcd 500.03, found 500.0.
[[(2R,3S,4R,5R)-5-[2-chloro-6-(cyclopropylamino)purin-9-yl]-3,4-dihydroxyoxolan-2-yl]methoxy-
hydroxyphosphoryl]methylphosphonic acid (9): The title compound was synthesized in similar fashion to Example 13
ACS Paragon Plus Environment

Page 25 of 46
Journal of Medicinal Chemistry
using cyclopropylamine in place of cyclopentylamine: 1H NMR (400 MHz, DMSO-d6) δ 8.54 (s, 1H), 8.42 (s, 1H), 5.86 (d, J =
5.8 Hz, 1H), 4.52 (t, J = 5.4 Hz, 1H), 4.28 – 4.03 (m, 4H), 2.97 (s, 1H), 2.25 (t, J = 20.5 Hz, 2H), 0.75 (s, 2H), 0.64 (s, 3H). ESI
MS [M+H]+ for C14H20ClN5O9P2, calcd 500.1, found 500.1
1
2
3
4
5
6
7
8
9
[[(2R,3R,4S,5R)-5-[2-chloro-6-(propan-2-ylamino)purin-9-yl]-4-fluoro-3-hydroxyoxolan-2-yl]methoxy-
hydroxyphosphoryl]methylphosphonic acid (10): The title compound was synthesized in similar fashion to Example 16:
1H NMR (400 MHz, DMSO-d6) δ 8.27 (m, 2H), 6.37 (d, J = 13.9 Hz, 1H), 5.28 (brs, 2H), 5.25 (d, J = 52.1 Hz, 1H), 4.98 (brs,
1H), 4.51 (d, J = 18.3 Hz, 1H), 4.35 (sept, J = 7.9 Hz, 1H), 4.19 (m, 2H), 4.04 (m, 1H), 2.26 (t, J = 20 Hz, 2H), 1.21 (dd, J = 6.6,
2.1 Hz, 6H). ESI MS [M-H]- for C14H21ClFN5O8P2, calcd 502.7, found 502.2.
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
[[(2R,3S,4R,5R)-5-[2-chloro-6-(propan-2-ylamino)purin-9-yl]-3,4-dihydroxyoxolan-2-yl]methoxy-
hydroxyphosphoryl]methylphosphonic acid (11): The title compound was synthesized in similar fashion to Example 13
using isopropylamine in place of cyclopentylamine: 1H NMR (400 MHz, DMSO-d6) δ 8.40 (s, 1H), 8.23 (d, J = 8.1 Hz, 1H),
5.85 (d, J = 5.9 Hz, 1H), 4.51 (t, J = 5.5 Hz, 1H), 4.36 (s, 1H), 4.24 – 4.03 (m, 4H), 2.25 (t, J = 20.5 Hz, 2H), 1.21 (dd, J = 6.6,
2.0 Hz, 5H). ESI MS [M+H]+ for C14H22ClN5O9P2, calcd 502.1, found 502.1
[[(2R,3R,4S,5R)-5-[2-chloro-6-(cyclopentylamino)purin-9-yl]-4-fluoro-3-hydroxyoxolan-2-yl]methoxy-
hydroxyphosphoryl]methylphosphonic acid (12): The title compound was synthesized in similar fashion to Example 16:
1H NMR (400 MHz, DMSO-d6) δ 8.41 (d, J = 7.8 Hz, 1H), 8.27 (s, 1H), 6.37 (dd, J = 14.4, 4.6 Hz, 1H), 5.25 (dt, J = 52.4, 4.3 Hz,
1H), 4.55 – 4.37 (m, 2H), 4.19 (t, J = 6.1 Hz, 2H), 4.03 (q, J = 5.1 Hz, 1H), 2.26 (t, J = 20.5 Hz, 2H), 1.93 (s, 2H), 1.64 (d, J =
62.5 Hz, 6H). ESI MS [M+H]+ for C16H23ClFN5O8P2, calcd 530.1, found 530.2
[[(2R,3S,4R,5R)-5-[2-chloro-6-(cyclopentylamino)purin-9-yl]-3,4-dihydroxyoxolan-2-yl]methoxy-
hydroxyphosphoryl]methylphosphonic acid (13): A mixture of 2,6-dichloropurine riboside 70 (321 mg, 1 mmol),
cyclopentylamine (103 µL, 1.05 mmol, 1.05 equiv.), and triethylamine (146 µL, 1.05 mmol, 1.05 equiv.) in anhydrous EtOH
(3 mL) was stirred at 60 °C for overnight. Reaction mixture was evaporated and the crude product was used in the next
step without purification. ESI MS [M+H]+ for C15H21ClN5O4, calcd 370.8, found 370.2.
The above product (370 mg, 1 mmol) was dissolved in trimethyl phosphate (5 mL) and cooled to 0 °C (ice bath), then a
cold solution of methylenebis(phosphonic dichloride) (1.25 g, 5 mmol, 5 equiv.) in trimethyl phosphate (2 mL) was added
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 46
dropwise. The reaction mixture was stirred at 0 °C for 3 h, and was then carefully quenched with 0.5 M triethylammonium
bicarbonate solution (7 mL) and stirred at 0 °C for 15 min, and then 2 h at room temperature. The reaction mixture was
purified by reverse phase HPLC (C18 column, 0 to 30% gradient of MeCN and water with 0.1% TFA) to give the product as
1
2
3
4
5
6
7
8
1
a white solid in 28% yield (181 mg): H NMR (400 MHz, DMSO) δ 8.45 – 8.32 (m, 2H), 5.85 (d, J = 5.5 Hz, 1H), 4.55 – 4.36
(m, 2H), 4.23 – 4.07 (m, 4H), 2.26 (t, J = 20.5 Hz, 2H), 2.04 – 1.85 (m, 2H), 1.77 – 1.46 (m, 6H). ESI MS [M+H]+ for
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
C16H25ClN5O9P2, calcd 528.8, found 528.1.
[[(2R,3R,4S,5R)-5-[2-chloro-6-[cyclopentyl(methyl)amino]purin-9-yl]-4-fluoro-3-hydroxyoxolan-2-yl]methoxy-
hydroxyphosphoryl]methylphosphonic acid (14): This compound was obtained similar fashion to Example 16 using N-
1
methylcyclopentylamine in place of benzylamine. H NMR (400 MHz, DMSO-d6) δ 8.7 (brs, 2H), 8.30 (d, J = 2.1 Hz, 1H),
6.40 (dd, J = 14.3, 4.6 Hz, 1H), 6.09 (brs, 1H), 5.25 (dt, J = 52.5, 4.3 Hz, 1H), 4.53-4.43 (m, 1H), 4.23-4.14 (m, 2H), 4.09 –
3.98 (m, 1H), 2.28 (dd, J = 20.5 Hz, J = 20.5 Hz, 2H), 2.5 (s, 3H), 1.96 – 1.44 (m, 9H). ESI MS [M+H]+ for C17H25ClFN5O8P2,
calcd 544.8, found: 544.2
[[(2R,3S,4R,5R)-5-[2-chloro-6-[cyclopentyl(methyl)amino]purin-9-yl]-3,4-dihydroxyoxolan-2-yl]methoxy-
hydroxyphosphoryl]methylphosphonic acid (15): The title compound was synthesized in similar fashion to Example 13
1
with N-methylcyclopentylamine in place of cyclopentylamine. H NMR (400 MHz, DMSO-d6) δ 8.42 (s, 1H), 5.88 (d, J = 5.9
Hz, 1H), 4.53 – 4.46 (m, 1H), 4.19 (dd, J = 5.0, 3.1 Hz, 1H), 4.15 – 4.06 (m, 3H), 3.17 (brs, 3H), 2.26 (t, J = 20.5 Hz, 2H), 1.94
– 1.53 (m, 9H). ESI MS [M+H]+ for C17H27ClN5O9P2, calcd 542.1, found 542.2.
[[(2R,3S,4R,5R)-5-[6-(benzylamino)-2-chloropurin-9-yl]-3,4-dihydroxyoxolan-2-yl]methoxy-
hydroxyphosphoryl]methylphosphonic acid (17): The title compound was synthesized in similar fashion to Example 13.
1H NMR (400 MHz, DMSO) δ 8.95 (t, J = 6.2 Hz, 1H), 8.44 (s, 1H), 7.46 – 7.14 (m, 5H), 5.86 (d, J = 5.8 Hz, 1H), 4.75 – 4.58
(m, 2H), 4.52 (t, J = 5.4 Hz, 1H), 4.26 – 3.97 (m, 4H), 2.26 (t, J = 20.5 Hz, 2H). ESI MS [M-H]- for C18H21ClN5O9P2, calcd 548.8,
found 548.2.
[[(2R,3R,4S,5R)-5-[2-chloro-6-[[(1R)-1-phenylethyl]amino]purin-9-yl]-4-fluoro-3-hydroxyoxolan-2-yl]methoxy-
hydroxyphosphoryl]methylphosphonic acid (18): The title compound was synthesized in similar fashion to Example 16.
1H NMR (400 MHz, DMSO-d6) δ 8.90 (d, J = 8.3 Hz, 1H), 8.28 (s, 1H), 7.42 (d, J = 7.6 Hz, 2H), 7.29 (t, J = 7.5 Hz, 2H), 7.19 (bs,
ACS Paragon Plus Environment

Page 27 of 46
Journal of Medicinal Chemistry
1H), 6.34 (dd, J = 14.8, 4.4 Hz, 1H), 5.39 (bs, 1H), 5.23 (d, J = 52.6 Hz, 1H), 4.49 (d, J = 19.2 Hz, 2H), 4.17 (bs, 2H), 4.01 (d, J
= 5.1 Hz, 1H), 2.24 (t, J = 20.6 Hz, 2H), 1.52 (d, J = 7.0 Hz, 3H). ESI MS [M+H]+ for C19H24ClFN5O8P2, calcd 566.1, found 566.1
1
2
3
4
5
6
7
8
[[(2R,3S,4R,5R)-5-[2-chloro-6-[[(1R)-1-phenylethyl]amino]purin-9-yl]-3,4-dihydroxyoxolan-2-yl]methoxy-
hydroxyphosphoryl]methylphosphonic acid (19): The title compound was synthesized in similar fashion to Example 13.
1H NMR (400 MHz, DMSO) δ 8.94 – 8.80 (m, 1H), 8.48 – 8.37 (m, 1H), 7.49 – 7.15 (m, 5H), 5.84 (d, J = 5.8 Hz, 1H), 5.41 (q,
J = 7.0 Hz, 1H), 4.57 – 4.44 (m, 1H), 4.23 – 4.01 (m, 4H), 2.26 (t, J = 20.5 Hz, 2H), 1.54 (d, J = 7.0 Hz, 3H). ESI MS [M+H]+ for
C19H25ClN5O9P2, calcd 564.1, found 564.1.
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
[[(2R,3R,4S,5R)-5-[2-chloro-6-[[(1S)-1-phenylethyl]amino]purin-9-yl]-4-fluoro-3-hydroxyoxolan-2-yl]methoxy-
hydroxyphosphoryl]methylphosphonic acid (20): The title compound was synthesized in similar fashion to Example 16.
1H NMR (400 MHz, DMSO-d6) δ 8.91 (d, J = 8.3 Hz, 1H), 8.28 (s, 1H), 7.41 (d, J = 7.5 Hz, 2H), 7.29 (t, J = 7.6 Hz, 2H), 7.20 (d,
J = 7.6 Hz, 1H), 6.34 (d, J = 14.1 Hz, 1H), 5.39 (bs, 1H), 5.21 (d, J = 52.5 Hz, 1H), 4.47 (d, J = 18.3 Hz, 2H), 4.17 (s, 2H), 4.01
(s, 1H), 2.24 (t, J = 20.6 Hz, 2H), 1.52 (d, J = 7.1 Hz, 3H). ESI MS [M+H]+ for C19H24ClFN5O8P2, calcd 566.1, found 566.1
[[(2R,3S,4R,5R)-5-[2-chloro-6-[[(1S)-1-phenylethyl]amino]purin-9-yl]-3,4-dihydroxyoxolan-2-yl]methoxy-
hydroxyphosphoryl]methylphosphonic acid (21): The title compound was synthesized in similar fashion to Example 13.
1H NMR (400 MHz, DMSO) δ 8.94 – 8.78 (m, 1H), 8.48 – 8.38 (m, 1H), 7.47 – 7.13 (m, 5H), 5.84 (d, J = 5.8 Hz, 1H), 5.41 (q,
J = 7.1 Hz, 1H), 4.54 – 4.46 (m, 1H), 4.22 – 4.05 (m, 4H), 2.26 (t, J = 20.5 Hz, 2H), 1.54 (d, J = 7.1 Hz, 3H). ESI MS [M+H]+ for
C19H25ClN5O9P2, calcd 564.1, found 564.2.
[[(2R,3R,4S,5R)-5-[2-chloro-6-[(2-chlorophenyl)methylamino]purin-9-yl]-4-fluoro-3-hydroxyoxolan-2-yl]methoxy-
hydroxyphosphoryl]methylphosphonic acid (22): The title compound was synthesized in similar fashion to Example 16.
1H NMR (400 MHz, DMSO-d6) δ 8.99 (t, J = 6.1 Hz, 1H), 8.35 (s, 1H), 7.47 (dd, J = 6.0, 3.3 Hz, 1H), 7.35 – 7.22 (m, 3H), 6.40
(dd, J = 14.2, 4.6 Hz, 1H), 5.27 (dt, J = 52.4, 4.3 Hz, 1H), 4.73 (d, J = 5.2 Hz, 2H), 4.52 (d, J = 18.5 Hz, 1H), 4.20 (t, J = 6.2 Hz,
2H), 4.05 (q, J = 5.1 Hz, 1H), 2.27 (t, J = 20.5 Hz, 2H). ESI MS [M+H]+ for C18H20Cl2FN5O8P2, calcd 586.0, found 586.1
[[(2R,3R,4S,5R)-5-[2-chloro-6-[(3-chlorophenyl)methylamino]purin-9-yl]-4-fluoro-3-hydroxyoxolan-2-yl]methoxy-
hydroxyphosphoryl]methylphosphonic acid (23): The title compound was synthesized in similar fashion to Example 16.
1H NMR (400 MHz, DMSO-d6) δ 9.03 (t, J = 6.2 Hz, 1H), 8.33 (d, J = 2.2 Hz, 1H), 7.45 – 7.27 (m, 4H), 6.39 (dd, J = 14.4, 4.6
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 46
Hz, 1H), 5.26 (dt, J = 52.4, 4.2 Hz, 1H), 4.74 – 4.58 (m, 2H), 4.51 (dt, J = 18.5, 4.6 Hz, 1H), 4.20 (t, J = 6.1 Hz, 2H), 4.04 (q, J =
5.1 Hz, 1H), 2.27 (t, J = 20.5 Hz, 2H). ESI MS [M-H]- for C18H19Cl2FN5O8P2, calcd 584.0, found 584.0
1
2
3
4
5
6
7
8
[[(2R,3R,4S,5R)-5-[2-chloro-6-[(4-chlorophenyl)methylamino]purin-9-yl]-4-fluoro-3-hydroxyoxolan-2-yl]methoxy-
hydroxyphosphoryl]methylphosphonic acid (24): The title compound was synthesized in similar fashion to Example 16.
1H NMR (400 MHz, DMSO-d6) δ 9.01 (t, J = 6.2 Hz, 1H), 8.32 (d, J = 2.1 Hz, 1H), 7.64 – 7.08 (m, 4H), 6.38 (dd, J = 14.3, 4.6
Hz, 1H), 5.26 (dt, J = 52.5, 4.3 Hz, 1H), 4.64 (q, J = 7.3, 5.4 Hz, 2H), 4.51 (dt, J = 18.7, 4.6 Hz, 1H), 4.28 – 4.11 (m, 2H), 4.04
(q, J = 5.1 Hz, 1H), 2.27 (t, J = 20.5 Hz, 2H). ESI MS [M-H]- for C18H20Cl2FN5O8P2, calcd 584.0, found 584.1.
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
[[(2R,3R,4S,5R)-5-(2-chloro-6-pyrrolidin-1-ylpurin-9-yl)-4-fluoro-3-hydroxyoxolan-2-yl]methoxy-
hydroxyphosphoryl]methylphosphonic acid (25): The title compound was synthesized in similar fashion to Example 16.
1H NMR (400 MHz, DMSO-d6) δ 8.27 (d, J = 2.1 Hz, 1H), 6.39 (dd, J = 14.1, 4.7 Hz, 1H), 5.26 (dt, J = 52.5, 4.3 Hz, 1H), 4.50
(dt, J = 18.5, 4.6 Hz, 1H), 4.19 (t, J = 5.9 Hz, 2H), 4.05 (q, J = 5.3, 4.1 Hz, 3H), 3.60 (t, J = 6.8 Hz, 2H), 2.27 (t, J = 20.5 Hz, 2H),
2.01 (p, J = 6.7 Hz, 2H), 1.92 (q, J = 6.7 Hz, 2H).ESI MS [M+H]+ for C15H21ClFN5O8P2, calcd 516.1, found 516.1.
[[(2R,3R,4S,5R)-5-[2-chloro-6-[[(3S)-oxolan-3-yl]amino]purin-9-yl]-4-fluoro-3-hydroxyoxolan-2-yl]methoxy-
hydroxyphosphoryl]methylphosphonic acid (26): The title compound was synthesized in similar fashion to Example 16.
1H NMR (400 MHz, DMSO-d6) δ 8.66 (d, J = 6.3 Hz, 1H), 8.31 (s, 1H), 6.38 (dd, J = 14.3, 4.6 Hz, 1H), 5.26 (dt, J = 52.4, 4.2 Hz,
1H), 4.62 (d, J = 12.0 Hz, 1H), 4.58 – 4.35 (m, 1H), 4.19 (t, J = 6.1 Hz, 2H), 4.04 (q, J = 5.0 Hz, 1H), 3.89 (dt, J = 15.3, 7.9 Hz,
2H), 3.78 – 3.68 (m, 1H), 3.61 (dd, J = 8.9, 4.4 Hz, 1H), 2.43 – 1.87 (m, 4H). ESI MS [M-H]- for C15H21ClFN5O9P2, calcd. 530.1,
found 530.1.
[[(2R,3R,4S,5R)-5-[2-chloro-6-(pyridin-4-ylmethylamino)purin-9-yl]-4-fluoro-3-hydroxyoxolan-2-yl]methoxy-
hydroxyphosphoryl]methylphosphonic acid (27): The title compound was synthesized in similar fashion to Example 16.
1H NMR (400 MHz, DMSO-d6) δ 9.13 (s, 1H), 8.66 (d, J = 5.7 Hz, 2H), 8.37 (s, 1H), 7.65 (d, J = 5.6 Hz, 2H), 6.40 (dd, J = 14.0,
4.6 Hz, 1H), 5.40 – 5.08 (m, 1H), 4.80 (d, J = 6.1 Hz, 2H), 4.53 (d, J = 18.3 Hz, 1H), 4.19 (s, 2H), 4.04 (d, J = 5.2 Hz, 1H), 2.25
(t, J = 20.4 Hz, 2H). ESI MS [M+H]+ for C17H20ClFN6O8P2, calcd 553.1, found 553.2
[[(2R,3R,4S,5R)-5-[2-chloro-6-(1,3-dihydroisoindol-2-yl)purin-9-yl]-4-fluoro-3-hydroxyoxolan-2-yl]methoxy-
hydroxyphosphoryl]methylphosphonic acid (28): The title compound was synthesized in similar fashion to Example 16.
ACS Paragon Plus Environment

Page 29 of 46
Journal of Medicinal Chemistry
1H NMR (400 MHz, DMSO-d6) δ 8.37 (d, J = 2.1 Hz, 1H), 7.48 (dt, J = 9.9, 4.7 Hz, 2H), 7.43 – 7.28 (m, 2H), 6.44 (dd, J = 13.8,
4.7 Hz, 1H), 5.41 (s, 2H), 5.29 (dt, J = 52.6, 4.4 Hz, 1H), 4.98 (s, 2H), 4.54 (dt, J = 18.7, 4.7 Hz, 1H), 4.21 (t, J = 5.9 Hz, 2H),
4.05 (q, J = 4.9 Hz, 1H), 2.27 (t, J = 20.5 Hz, 2H). ESI MS [M+H]+ for C19H21ClFN5O8P2, calcd 564.1, found 564.1
1
2
3
4
5
6
7
8
[[(2R,3R,4S,5R)-5-(2-chloro-6-phenylmethoxypurin-9-yl)-4-fluoro-3-hydroxyoxolan-2-yl]methoxy-
9
hydroxyphosphoryl]methylphosphonic acid (29): Under a nitrogen atmosphere, sodium hydride (90 mg, 2.26 mmol, 1.2
equiv., 60% in oil) and benzyl alcohol (10 mL) were stirred at r.t. for 15 min. Compound 62 (1.00 g, 1.88 mmol) was added
and the mixture stirred at r.t. for 2 h. The reaction mixture was purified directly by column chromatography (0-10% MeOH
in CH2Cl2) to afford 69 as a white solid (721 mg, 97%). ESI MS [M+H]+ for C17H17ClFN4O4, calcd 395.1, found 395.1.
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
Intermediate 69 (197 mg, 0.5 mmol) was dissolved in trimethyl phosphate (2.5 mL) and cooled to 0 °C. A solution of
methylenebis(phosphonic dichloride) (624 mg, 2.5 mmol, 5 equiv.) in trimethyl phosphate (1.5 mL) was added dropwise.
The reaction mixture was stirred at 0 °C for 3 h and then carefully quenched at -20 °C with 0.5 M triethylammonium
bicarbonate solution (3.6 mL). The mixture was stirred at -20 °C for 15 min, then stirred at 0 °C for 15 min, then stirred at
r.t. for 15 min. The mixture was washed with EtOAc (10 mL) three times. The aqueous layer was purified directly by reverse
phase HPLC (C18 column, 0 to 50% gradient of MeCN and water with 0.1% TFA) to afford the desired product as a white
solid (40.2 mg, 15%): 1H NMR (400 MHz, DMSO-d6) δ 8.55 (d, J = 2.1 Hz, 1H), 7.57 – 7.51 (m, 2H), 7.46 – 7.35 (m, 3H), 6.49
(dd, J = 13.6, 4.7 Hz, 1H), 5.61 (s, 2H), 5.30 (dt, J = 52.4, 4.4 Hz, 1H), 4.53 (dt, J = 18.6, 4.7 Hz, 1H), 4.21 (t, J = 6.0 Hz, 2H),
4.06 (q, J = 5.0 Hz, 1H), 2.27 (t, J = 20.6 Hz, 2H). ESI MS [M-H]- for C18H19ClFN4O9P2, calcd 551.0, found 551.2.
[[(2R,3R,4S,5R)-5-[6-(benzylamino)-2-methylpurin-9-yl]-4-fluoro-3-hydroxyoxolan-2-yl]methoxy-
hydroxyphosphoryl]methylphosphonic acid (30): The title compound was synthesized in similar fashion to Example 16
using 6-chloro-2-methylpurine in place of 2,6-dichloropurine:1H NMR (400 MHz, DMSO-d6) δ 8.51 (s, 1H), 8.23 (s, 1H), 7.44
– 7.19 (m, 5H), 6.44 (dd, J = 15.0, 4.6 Hz, 1H), 5.41 – 5.13 (m, 1H), 4.72 (s, 2H), 4.53 (dd, J = 18.4, 4.7 Hz, 1H), 4.19 (t, J =
6.1 Hz, 2H), 4.04 (t, J = 5.1 Hz, 1H), 2.46 (s, 3H), 2.26 (t, J = 20.5 Hz, 2H). ESI MS [M+H]+ for C19H24FN5O8P2, calcd 532.1,
found 532.2.
[[(2R,3S,4R,5R)-5-[6-(benzylamino)-2-methylpurin-9-yl]-3,4-dihydroxyoxolan-2-yl]methoxy-
hydroxyphosphoryl]methylphosphonic acid (31): To a nitrogen purged reaction mixture of the N6-benzyl-2-iodopurine
riboside 72 (1.03 g, 1.7 mmol) and tetramethyltin (470 µL, 3.34 mmol) in NMP (10 mL) was added Pd(PPh3)4 (196 mg, 0.17
ACS Paragon Plus Environment