A Structure-Activity Relationship Study of Bitopic N 6-Substituted Adenosine Derivatives as Biased Adenosine A1 Receptor Agonists

Article abstract of DOI:10.1021/acs.jmedchem.8b00047

The adenosine A₁ receptor (A₁AR) is a potential novel therapeutic target for myocardial ischemia-reperfusion injury. However, to date, clinical translation of prototypical A₁AR agonists has been hindered due to dose limiting adverse effects. Recently, we demonstrated that the biased bitopic agonist 1, consisting of an adenosine pharmacophore linked to an allosteric moiety, could stimulate cardioprotective A₁AR signaling in the absence of unwanted bradycardia. Therefore, this study aimed to investigate the structure-activity relationship of compound 1 biased agonism. A series of novel derivatives of 1 were synthesized and pharmacologically profiled. Modifications were made to the orthosteric adenosine pharmacophore, linker, and allosteric 2-amino-3-benzoylthiophene pharmacophore to probe the structure-activity relationships, particularly in terms of biased signaling, as well as A₁AR activity and subtype selectivity. Collectively, our findings demonstrate that the allosteric moiety, particularly the 4-(trifluoromethyl)phenyl substituent of the thiophene scaffold, is important in conferring bitopic ligand bias at the A₁AR.

Full text of DOI:10.1021/acs.jmedchem.8b00047

Subscriber access provided by UNIV OF DURHAM
Article
A Structure-Activity Relationship Study of Bitopic N6-Substituted
Adenosine Derivatives as Biased Adenosine A1 Receptor Agonists
Luigi Aurelio, Jo-Anne Baltos, Leigh Ford, Anh T.N. Nguyen, Manuela Jörg, Shane M. Devine,
Celine Valant, Paul J. White, Arthur Christopoulos, Lauren T. May, and Peter J. Scammells
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.8b00047 • Publication Date (Web): 15 Feb 2018
Downloaded from http://pubs.acs.org on February 16, 2018
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.
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 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
6
A Structure-Activity Relationship Study of Bitopic N -Substituted
Adenosine Derivatives as Biased Adenosine A Receptor Agonists
1
†
,§
‡,§
†
‡
†
Luigi Aurelio, JoꢀAnne Baltos, Leigh Ford, Anh T. N. Nguyen, Manuela Jörg,
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
†
‡
‡
‡
Shane M. Devine, Celine Valant, Paul J. White, Arthur Christopoulos,
‡
,*
†,*
Lauren T. May, Peter J. Scammells
†
‡
§
Medicinal Chemistry, Monash Institute of Pharmaceutical Sciences; Monash University,
Parkville, VIC 3052, Australia
Drug Discovery Biology, Monash Institute of Pharmaceutical Sciences & Department of
Pharmacology, Monash University, Parkville, VIC 3052, Australia
These authors contributed equally to this manuscript
1 1
ABSTRACT: The adenosine A receptor (A AR) is a potential novel therapeutic target for
myocardial ischemiaꢀreperfusion injury. However, to date, clinical translation of prototypical
A AR agonists has been hindered due to dose limiting adverse effects. Recently, we demonstrated
1
that the biased bitopic agonist 1, consisting of an adenosine pharmacophore linked to an allosteric
1
moiety, could stimulate cardioprotective A AR signalling in the absence of unwanted bradycardia.
Therefore, this study aimed to investigate the structureꢀactivity relationship of compound 1 biased
agonism. A series of novel derivatives of 1 were synthesised and pharmacologically profiled.
Modifications were made to the orthosteric adenosine pharmacophore, linker and allosteric 2ꢀ
aminoꢀ3ꢀbenzoylthiophene pharmacophore to probe the structureꢀactivity relationships, particularly
in terms of biased signalling, as well as A
1
AR activity and subtype selectivity. Collectively, our
findings demonstrate that the allosteric moiety, particularly the 4ꢀ(trifluoromethyl)phenyl
1
substituent of the thiophene scaffold, is important in conferring bitopic ligand bias at the A AR.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ꢀ
INTRODUCTION
The adenosine receptor family consists of four subtypes, namely the A
1
AR, A2AAR, A2BAR and
1,2
A
3
AR, which are part of the G proteinꢀcoupled receptor (GPCR) superfamily.
The A
1
AR is an
3
ꢀ6
important therapeutic target, known to stimulate protective effects in heart tissue, particularly in
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
7
,8
models of ischemiaꢀreperfusion injury (IRI).
However, there is a high attrition rate in the
transition of A AR agonists into clinic due to dose limiting side effects such as hypotension,
1
9
ꢀ11
bradycardia and atrioventricular block.
therapeutic and adverse effects are mediated by the A
subtype selective agonists, A AR biased agonists have the potential to overcome these commonly
observed limitations. Biased agonists can stimulate distinct signalling profiles to that of
Therefore, the primary obstacle is that both the
12,13
1
AR.
1
In contrast to prototypical A AR
1
14ꢀ16
prototypical agonists through stabilizing a different spectrum of receptor conformations.
As
such, GPCR biased agonism provides the opportunity to design compounds that selectivity
stimulate the pathway associated with the therapeutic effect, whilst avoiding signalling associated
with onꢀtarget adverse effects.
Our group has previously reported the identification of VCP746 (1) (Figure 1), a rationally
designed adenosine receptor bivalent ligand that has a bitopic mechanism of action and biased
1
3
profile at the A
1
AR. Bitopic ligands are bivalent molecules that interact simultaneously with an
17,18
orthosteric site and allosteric binding site within a GPCR monomer.
Compound 1 consists of
the orthosteric agonist adenosine and a biased allosteric thiophene pharmacophore, which are
tethered together by an aromatic linker unit and a 6ꢀcarbon alkyl linker (Figure 1). Pharmacological
evaluation of 1 revealed significant bias away from calcium mobilization, compared to ERK1/2
phosphorylation, Akt phosphorylation and inhibition of cAMP accumulation, relative to the
1
9
reference agonist NECA. Furthermore, we were able to demonstrate the compounds ability to
1
protect against ischemia in native A ARꢀexpressing cardiomyoblasts without effects on the rat atrial
13
heart rate.
In summary, our proof of concept studies have shown a link between the biased
pharmacological profile of compound 1 to its ability to promote onꢀtarget efficacy without onꢀtarget
ACS Paragon Plus Environment

Page 3 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
adverse effects. This work embodied an important step in the quest of rationally identifying better
1
drugs that act at the A AR.
Orthosteric Pharmacophore
Linker
NH2
ortho /
meta /
Para
S
O
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
NH2
O
S
O
HN
N
n
N
H
H
N
N
N
HN
N
O
N
N
HO
F3C
O
N
F3C
N
X
Y
O
HO
OH
1 X = H, Y = CH OH (VCP746)
2
2
3
4
X = NH , Y = CH OH
2 2
HO
OH
Homologation Strategy Thiophene Walk
X = Cl, Y = CH OH
2
5 n= 5
6 n= 7
7 meta thiophene
8 ortho thiophene
X = H, Y = C(O)NHEt
Allosteric Pharmacophore
R2
2
-Amino Acetylation
2
S
9 R = NHC(O)CH
3
R3
H
R4
Deletion Strategy
N
2
3
4
1
0 R = H, R = C(O)Ph, R = 3-CF Ph
HN
3
11 R2 = NH , R = H, R = 3-CF Ph
3
4
2
3
O
N
N
2
3
4
12 R = NH , R = C(O)Ph, R = H
2
2
3
4
13 R = H, R = H, R = 3-CF Ph
HO
N
3
N
2
3
4
O
14 R = H, R = C(O)Ph, R = H
2
3
4
15 R = NH , R = R = H
2
2
3
4
16 R = R = R = H
HO
OH
1
Figure 1. Chemical structure of the biased bitopic A AR ligand 1 and an overview of the structural
changes made as part of the SAR study. The colours represent the different components of the
bitopic ligand; namely, the orthosteric pharmacophore (blue), linker (black) and the allosteric
pharmacophore (red).
The aim of the current study was to understand the structureꢀactivity relationships underlying
1
the physiologically relevant A AR biased agonism mediated by compound 1. Herein, we describe
the synthesis and pharmacological evaluation of a series of compounds that derived from 1. We
have made structural changes to the orthosteric pharmacophore (2–4), the linker (5–8) and the
allosteric pharmacophore (9–16) (Figure 1). Modification to the 2ꢀposition of the purine scaffold
and the 5’ꢀposition of the sugar moiety have been shown to affect the subtype selectivity of
adenosine agonists. Therefore, we synthesized the 2ꢀamino (2) and 2ꢀchloro (3) substituted analogs
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
of 1 as well as the 5’ꢀethylcarboxamide derivative (4). The linker has been modified to probe the
effects of subtle changes in length (i.e. one fewer or one additional methylene unit, 5 and 6
respectively) as well as its point of attachment to the thiophene (meta / ortho, 7 and 8 respectively).
Finally, the allosteric moiety has been altered either by acetylation of the amino group (9) or a
deletion strategy removing one (10-12), two (13-15) or all three substituents (16) of the thiophene
ring.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ꢀ RESULTS AND DISCUSSION
Chemistry. The synthetic pathway used for the preparation of the bitopic analogs was based on
the general approach shown in Scheme 1. First, the respective orthosteric pharmacophore (17a–d)
was functionalized with the diaminoalkyl spacer (18a–c), which then was coupled with the
respective carboxylic acid functionalized allosteric portion under standard coupling conditions to
obtain the final bitopic analogs (1–16).
a
1
Scheme 1. Synthesis of the bitopic adenosine A receptor agonists
Cl
HN
N
n
NH2
N
N
N
N
N
functionalized allosteric
pharmacophore
H N
NH2
n
a
2
HO
N
N
X
X
O
Y
21a-d, 22a-c, 25, 27, 34a-b
O
HO
OH
HO
OH
1
7a X= H
18a n= 5
18b n= 6
18c n= 7
19a X= H, Y= CH OH, n= 5
2
17b X= NH2
1
19b X= H, Y= CH OH, n= 6
2
7c X= Cl
19c X= H, Y= CH OH, n= 7
2
1
9d X= NH , Y= CH OH, n= 6
2 2
1
9e X= Cl, Y= CH OH, n= 6
2
19f X= H, Y= C(O)NHEt, n= 6
b
N
O
N
allosteric
pharmacophore
N
O
N
HN
N
n
N
H
N
N
N
N
N
N
H
N
O
O
X
Y
O
bitopic ligands 2-16
1
7d
HO
OH
HO
OH
a
Reagents and conditions: (a) (i) Et
3
N, EtOH, reflux, 18ꢀ92% for 19a–e; (ii) diaminoalkane (18b),
N, DMF, rt, 6ꢀ90%.
Et N, EtOH, rt, product not isolated for 19f; (b) BOP, Et
3
3
ACS Paragon Plus Environment

Page 5 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Therefore, the respective diaminoalkanes (18a–c) were introduced via a substitution reaction
with the 6ꢀchloropurine riboside derivatives 17a–c in ethanol and triethylamine. The reactions were
stirred at reflux until complete conversion was observed, upon cooling to room temperature the
resultant precipitate was filtered to give intermediates 19a–d in good yields (54ꢀ92%); with the
exception of 19e which was obtained as a light yellowish oil in only 18% yield. Intermediate 19f
was synthesized using the same conditions but at room temperature and starting from
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(2S,3S,4R,5R)ꢀ5ꢀ(6ꢀ((1Hꢀbenzo[d][1,2,3]triazolꢀ1ꢀyl)oxy)ꢀ9Hꢀpurinꢀ9ꢀyl)ꢀNꢀethylꢀ3,4ꢀ
dihydroxytetrahydrofuranꢀ2ꢀcarboxamide (17d), which was prepared from the commercially
available inosine using a sixꢀstep synthetic route based on reported methods in the literature (see
2
0ꢀ22
Supporting Information).
The final bitopic ligands 2–16 were obtained by linking the orthosteric amine intermediates
9a–f and the prior prepared carboxylic acid functionalized allosteric moieties 21a–d, 22a–c, 25, 27
1
and 34a–b (Scheme 2–4). The formation of the amide bond was achieved using standard peptide
coupling conditions and afforded the bitopic ligands 2–16 in yields ranging between 6–90%.
The functionalized carboxylic acid allosteric moieties were synthesized via the routes
illustrated in Schemes 2–4. We have previously published the synthesis of the iodothiophenes 20
(Scheme 2), which can be achieved via either a twoꢀstep or oneꢀpot microwaveꢀassisted Gewald
23,24
procedure starting from the respective benzoylacetonitrile and acetophenone,
followed by the
1
3
acetylation of the amino group and iodination in the 5ꢀposition.
The carboxylic acid
functionalized allosteric moieties 21a–d were afforded in good yields by a Suzukiꢀcoupling reaction
with 20a or 20b and the respective ortho, meta or paraꢀsubstituted methoxycarbonylphenyl boronic
acids, followed by the simultaneous hydrolysis of the ester and deacetylation under alkaline
conditions. Acetylation of 21a with acetic anhydride gave the diacetylated intermediate (Oꢀand Nꢀ
acetylation), which was treated with lithium hydroxide to afford 22a in 98% yield. Alternatively,
21a and 21d were treated with tꢀbutylnitrite in DMF at 65 °C to remove the amino group and
provide compounds 22b and 22c in 91% and 71% yield, respectively.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
a
1
Scheme 2. Synthesis of the functionalized allosteric A AR pharmacophores 20a-d and 22a-c
HOOC
O
HOOC
I
S
S
S
NH
O
NH2
O
R'
O
(
a) or (b)
(c) or (d)
R
R
R
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
0a R= 3-CF Ph
21a 4-COOH, R= 3-CF Ph
22a R= 3-CF Ph, R'=NHC(O)CH
3
3
3
3
20b R= H
21b 3-COOH, R= 3-CF Ph
22b R= 3-CF Ph, R'= H
22c R= H, R'= H
3
3
2
2
1c 2-COOH, R= 3-CF Ph
1d 4-COOH, R= H
3
a
Reagents and conditions: (a) 3ꢀ or 4ꢀmethoxycarbonylphenyl boronic acid, Pd(PPh
PO , DMF, 70 °C; followed by NaOH, EtOH: H
methoxycarbonyl)phenyl)boronic acid, CsF, Pd(dppf)•DCM, dioxane, H
3
)
2
Cl
2
, 2 M
K
3
4
2
O 1:1, 40 °C, 37ꢀ89%; (b) (2ꢀ
O, 60 °C, followed by
2
O 1:1, rt, 31% of 21c; (c) 21a, acetic anhydride, 90 C, followed by LiOH, THF:
(
2
o
NaOH, EtOH: H
H
2
O 1:1, rt, 98% of 22a; (d) tꢀbutylnitrite, DMF, 65 °C, 91% of 22b (from 21a), 71% of 22c (from
21d).
A Suzukiꢀcoupling reaction of the commercially available starting materials 23a and 23b with
ꢀmethoxycarbonylphenyl boronic acid afforded intermediates 24a and 24b in 35% and 76% yield,
4
respectively (Scheme 3). Hydrolysis of 24b under mild conditions with lithium hydroxide gave the
carboxylic acid functionalized allosteric moieties 25 in modest yield (27%). Intermediate 24a
underwent a second Suzukiꢀcoupling reaction with (3ꢀ(trifluoromethyl)phenyl)boronic acid to
afford 26 in 64% yield, followed by the hydrolysis of the ester group with sodium hydroxide in
ethanol and water to give the carboxylic acid intermediate 27 in good yield (83%).
ACS Paragon Plus Environment

Page 7 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
1
Scheme 3. Synthesis of the functionalized allosteric A AR pharmacophores 25 and 27
O
O
H CO
3
RO
X
Y
S
(a)
S
(c)
S
Y
2
2
3a X= Br, Y=Br
3b X= I, Y= H
24a Y= Br
24b Y= H
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
CF3
(
b)
26 R= CH3
27 R= H
(d)
O
HO
S
2
5
a
Reagents and conditions: (a) (4ꢀ(methoxycarbonyl)phenyl)boronic acid, Pd(PPh
PO , DMF, 80 °C, 35% of 24a, 76% of 24b; (b) LiOH, EtOH: H
trifluoromethyl)phenyl)boronic acid, Pd(PPh Cl , 2 M K PO , DMF, 80 °C, 64%; (d) NaOH,
EtOH: H O 3:1, rt, 83%.
3
)
2
Cl
2
, 2 M
K
3
4
2
O 1:1, 27%; (c) (3ꢀ
(
3
)
2
2
3
4
2
The multistep synthesis of 34a and 34b (Scheme 4) started from the commercially available 5ꢀ
or 4ꢀbromothiopheneꢀ2ꢀcarboxylic acids 28a and 28b, respectively. First, the carboxylic acid
functional group was protected with a benzyl group to give 29a and b in quantitative yields. In the
case of 29b, a trifluoromethylphenyl substituent was introduced in the 4ꢀposition via a Suzukiꢀ
coupling reaction and the resulting product 29c was brominated in position 5 with Nꢀ
bromosuccinimide (NBS) to give 29d in 66% yield. The bromine intermediates 29a and 29d were
reacted with (4ꢀ(methoxycarbonyl)phenyl)boronic acid under Suzukiꢀcoupling conditions to afford
3
0a and 30b, subsequent debenzylation via hydrogenation with Pd/C gave intermediates 31a and
1b, respectively. Next, the carboxylic acid group was converted to an acyl azide with
3
diphenylphosphoryl azide (DPPA) in tertꢀbutanol, followed by an in situ Curtius rearrangement to
form the tertꢀbutyloxycarbonyl (Boc) protected intermediates 32a and 32b. These key
intermediates were then hydrolysed to the carboxylic acid analogues (33a and 33b), followed by the
final Bocꢀdeprotection under acidic conditions to form the hydrochloride salts 34a and 34b.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Finally, the carboxylic acid functionalized allosteric moieties 21a–d, 22a–c, 25, 27 and 34a–b
were linked to the orthosteric counterpart to form the final bitopic ligands 2–16, which were fully
characterized and assessed in our pharmacological assays.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
Scheme 4. Synthesis of the functionalized allosteric A
1
AR pharmacophores 34a–b
O
H CO
3
X
Y
X
S
O
S
O
(d)
S
O
(
a)
OH
OBn
OR
Y
Y
2
2
8a X= Br, Y= H
8b X= H, Y= Br
2
9a X= Br, Y= H
9b X= H, Y= Br
30a R= Bn, Y= H
0b R= Bn, Y= 2-CF Ph
3
3
2
2
2
(
b)
c)
(e)
9c X= H, Y= 2-CF Ph
3
3
1a R= H, Y= H
(
9d X= Br, Y= 2-CF Ph
31b R= H, Y= 2-CF Ph
3
3
(f)
O
O
O
HO
HO
H CO
3
S
S
(g)
S
(
h)
NH3 Cl
NHBoc
NHBoc
Y
Y
Y
3
3
4a Y= H
4b Y= 2-CF Ph
33a Y= H
33b Y= 2-CF Ph
32a Y= H
32b Y= 2-CF Ph
3
3
3
a
Reagents and conditions: (a) BnBr, TBAI, Cs CO , DMF, rt, 98%; (b) (3ꢀ
2
3
(trifluoromethyl)phenyl)boronic acid, CsF, Pd(dppf)•DCM, dioxane, H
2
O, 60 °C, 66%; (c) NBS,
.
AcOH, 60 °C, 100%; (d) (4ꢀ(methoxycarbonyl)phenyl)boronic acid, CsF, Pd(dppf) DCM, dioxane,
H O, 60 °C, 76ꢀ84%; (e) H , Pd/C, EtOAc, DMF or iPrOH, 82ꢀ89%; (f) tertꢀBuOH, DPPA, Et N,
2
2
3
2
rt, 69ꢀ71%; (g) 2 M NaOH, (dioxane), EtOH, H O, 85ꢀ98%; (h) 4 M HCl, dioxane, 87ꢀ100%.
1
Pharmacology. cAMP accumulation assays. The A AR preferentially couples to Gi/o proteins,
leading to a decrease in adenylyl cyclase activity and subsequent cAMP accumulation. As such, the
ability of NECA (a nonꢀselective adenosine receptor agonist), 1 and the bitopic derivatives (2–16)
to inhibit forskolinꢀstimulated cAMP accumulation was assessed in FlpInꢀCHO cells stably
1 1
expressing the human A AR (A ARꢀFlpInꢀCHO). Data were fitted to a threeꢀparameter Hill
equation and a reꢀparameterization of the operational model of agonism. The Hill equation
ACS Paragon Plus Environment

Page 9 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
provided estimation of agonist potency (pEC ) and maximal effect (EMAX), whereas the operational
5
0
A
model enabled quantification of the agonist transduction coefficient (Log(τ/K )), a composite value
2
5
of agonist affinity (K ) and efficacy (τ). The transduction coefficient is a useful parameter for
A
comparative pharmacology, particularly for agonists that have different maximal effects (EMAX), as
incorporating both agonist efficacy and affinity enables a more global measure of agonist activity.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
A
Subsequent normalization of Log(τ/K ) values to the reference agonist NECA generated
ꢀ
Log(τ/K ), a system independent measurement of agonist activity that was used for the
A
quantification of bias and subtype selectivity.
Modification to the orthosteric pharmacophore (i.e. the introduction of a 2ꢀamino, 2ꢀchloro or
5’ꢀNꢀethylcarboxamide moieties in 2, 3 and 4, respectively) resulted in a significant reduction of
Log(τ/K ) relative to the parent compound 1 (Table 1). In contrast, linker modifications had
A
differential effects. Decreasing the linker length by one methylene unit (5) had no significant
effect, while incorporation of an additional methylene linker (6) resulted in a significant increase in
A
the Log(τ/K ). Varying the point of attachment of the thiophene to the linker from a para to an
ortho substitution pattern (8) resulted in a significant decrease in the pEC50 and Log(τ/K ). A
A
number of modifications were also made to the allosteric pharmacophore. Neither acetylation or
removal of the amino substituent on the thiophene ring (9 and 10, respectively) nor individual or
dual removal of the benzoyl or trifluoromethylphenyl groups from the thiophene (11, 12 and 15)
had a significant effect on the pEC50 or Log(τ/K
A
A
). However, the Log(τ/K ) was significantly
reduced upon removal of the amino group together with one or both of the other thiophene
substituents (13, 14 and 16). All derivatives had a similar EMAX for cAMP inhibition, relative to the
nonꢀselective full agonist NECA (Table 1).
These structureꢀactivityꢀrelationships differ from what has been previously observed for ‘standꢀ
alone’ A
1
AR agonists and allosteric modulators. For example, the introduction of a 2ꢀchloro group
6
or 5ꢀNꢀethylcarboxamide to an N ꢀsubstituted adenosine has been shown to increase potency at the
26,27
A
1
AR.
Similarly, the removal of the 3ꢀbenzoyl moiety from a 2ꢀaminoꢀ3ꢀbenzoylthiophene
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
PAM abolishes activity. These differences in SAR profiles suggest that presence of the linker
induces a different binding mode for the bitopic molecules relative to the standꢀalone agonist and
allosteric modulator components.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ACS Paragon Plus Environment

Page 11 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
Table 1. A
1
AR Receptor Pharmacology: Inhibition of cAMP Accumulation and Stimulation of Intracellular Calcium Mobilization.
S 2
NH2
Orthosteric Pharmacophore
2-NH
3 2-Cl
5'-C(O)NHEt
Allosteric Pharmacophore
2-NHC(O)CH
10 des 2-NH2
11 des 3-C(O)Ph
12 des 4-PhCF3
2
2
9
3
H
3
HN (CH
N
2
)
n
N
O
4
4
N
N
O
Linker Modifications
5 n= 5
13 des 2-NH2 + des 3-C(O)Ph
2
HO 5'
3
F C
N
2 3
14 des 2-NH + des 4-PhCF
O
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
6
7
n= 7
meta thiophene
15 des 3-C(O)Ph + des 4-PhCF3
16 des 2-NH2 + des 3-C(O)Ph + des 4-PhCF3
HO
OH
8 ortho thiophene
Cpd
NECA 9.12
cAMP
Calcium
pEC50
E
Max
Log(
8.98 ± 0.09
7.35 ± 0.07
τ
/K )
n
10
3
4
4
4
3
3
3
3
3
3
3
3
3
3
3
3
pEC50
E
Max
Log(
τ
/K )
n
12
6
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
A
A
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
0.08
0.07
0.05****
0.11**
0.16
0.08
0.10
0.19
0.20****
0.11
0.08
0.13
0.11
0.08*
0.17
101
112
115
106
112
114
105
±
±
±
±
±
±
±
3
4
4
5
8
6
5
9.36 ± 0.11
6.91 ± 0.24
6.59 ± 0.38
6.87 ± 0.26
7.25 ± 0.40
7.13 ± 0.25
6.70 ± 0.45
7.03 ± 0.06
5.87 ± 0.05
6.05 ± 0.71
6.25 ± 0.48
6.96 ± 0.11
7.53 ± 0.06
7.11 ± 0.06
7.06 ± 0.11
7.46 ± 0.10
7.45 ± 0.08
54 ± 2
34 ± 5#
34 ± 9
20 ± 4
16 ± 5
31 ± 4#
28 ± 8#
48 ± 1
43 ± 1
20 ± 16
40 ± 17
42 ± 2
49 ± 1
49 ± 1
39 ± 2
47 ± 2
49 ± 1
9.31 ± 0.11
6.12 ± 0.34
6.15 ± 0.43
6.32 ± 0.21
6.33 ± 0.23
6.11 ± 0.42
5.58 ± 0.71
6.98 ± 0.05
5.74 ± 0.05
5.02 ± 0.43
5.74 ± 0.43
6.86 ± 0.06
7.43 ± 0.01*
7.03 ± 0.05
6.91 ± 0.03
7.40 ± 0.01*
7.38 ± 0.06*
1
2
3
4
5
6
7
8
9
1
1
1
1
1
7.37
5.87
6.65
7.01
7.02
7.64
6.98
5.89
6.86
6.96
7.42
7.38
6.77
6.86
7.17
6.63
6.01 ± 0.04****
6.70 ± 0.03**
6.80 ± 0.09**
6.98 ± 0.08
7.85 ± 0.13*
6.97 ± 0.23
5.89 ± 0.15****
6.97 ± 0.06
6.93 ± 0.11
7.44 ± 0.10
7.21 ± 0.04
6.64 ± 0.15***
6.86 ± 0.09*
7.04 ± 0.15
#
#
##
###
99
± 9
104
119
114
110
101
106
109
100
117
±
±
±
±
±
±
±
±
±
15
9
6
6
5
5
9
7
#
##
0
1
2
3
4
15
16
0.17
0.10**
6
6.69 ± 0.09**
Data represents the mean ± SEM of 3 to 12 separate experiments conducted in duplicate or triplicate. * P<0.05, ** P<0.01, *** P<0.001, ****
P<0.0001 derivative value (pEC50 or Log(τ/K
compound, 1, at the corresponding pathway. P<0.05, P<0.01, P<0.001, significantly different (oneꢀway analysis of variance, Dunnett’s postꢀhoc)
A
)) was significantly different (oneꢀway analysis of variance, Dunnett’s postꢀhoc) compared to the parent
#
##
###
19
when compared to the NECA EMAX at the corresponding pathway. Curves for parent compound, 1, were taken from Baltos et al.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Biased agonism at the A
1
1
AR. The influence of structural modifications on A AR biased
agonism was also investigated. We have previously demonstrated that in A ARꢀFlpInꢀCHO cells,
1
compound 1 has significant bias away from calcium mobilization, relative to the reference pathway
19
cAMP and the reference ligand, NECA.
The current study has quantitatively evaluated the
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
influence of various structural components of 1 on calcium mobilization (Table 1) and bias (Figure
A). Bias was quantified by normalizing agonist Log(τ/K ) values to the Log(τ/K ) value of the
reference agonist, NECA; subsequently, these values were normalized to the reference pathway,
2
A
A
19
cAMP accumulation, to generate Log(Bias Factor) values.
Modification to the orthosteric pharmacophore or linker had no significant effect on the pEC50
or Log(τ/K ) value, relative to 1, for A ARꢀmediated calcium mobilization (Table 1). In contrast,
A
1
single deletion of the 4ꢀ(trifluoromethyl)phenyl, dual deletion of the 4ꢀ(trifluoromethyl)phenyl in
conjunction with the benzoyl or triple deletion of all three substituents from the allosteric moiety
resulted in a significant increase in the Log(τ/K ) value for calcium mobilization (Table 1). These
A
findings suggest the 4ꢀ(trifluoromethyl)phenyl substituent on the allosteric pharmacophore impedes
A ARꢀmediated calcium mobilization.
1
Addition of a 2ꢀamino group to the purine scaffold with the orthosteric pharmacophore (2)
resulted in a significant loss of bias relative to 1, such that this compound was nonꢀbiased (Figure
2
A). The change in bias resulted from a decrease in the cAMP ꢀLog(τ/K ) with little reduction in
A
A
the calcium ꢀLog(τ/K ) (Figure 2B). Interestingly, increasing the linker length by one additional
carbon, as in compound 6, resulted in approximately 10ꢀfold increase in bias away from calcium
mobilization, relative to compound 1, however this effect did not reach significance (Figure 2A).
The change in the compound 6 bias factor arises due to an increase in the cAMP ꢀLog(τ/K ) but a
A
decrease in the calcium ꢀLog(τ/K
A
) (Figure 2B). We have previously shown that bitopic linker
13
length exhibits a bellꢀshaped relationship with ligand affinity and potency.
Accordingly, the
effects of the linker length on bias may reflect an improvement or weakening in the engagement of
ACS Paragon Plus Environment

Page 13 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2
8ꢀ30
the allosteric or orthosteric pharmacophore with their corresponding binding pockets,
thereby
enhancing or reducing biased signaling, respectively. Changing the position of the thiophene
scaffold from the para to the metaꢀposition (7) significantly decreased the degree of bias,
suggesting that the orientation of the allosteric pharmacophore is also important for biased signaling
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(Figure 2A).
Figure 2. (A) Bias factors for 1 and derivatives at the A
relative to the reference ligand NECA and the reference pathway cAMP. Agonist Log(τ/K
were normalized to the Log(τ/K ) of the reference ligand, NECA, at each respective pathway to
generate ꢀLog(τ/K ) values. These were then normalized to the cAMP pathway, to generate the
1
AR for the calcium mobilization pathway,
A
) values
A
A
Log(Bias Factor). *P<0.05, **P<0.01, Oneꢀway ANOVA Dunnett’s postꢀhoc analysis when
compared to the Log(Bias Factor) of 1. ^P<0.05, ^^P<0.01, unpaired tꢀtest comparing the calcium
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ꢀLog(τ/K
A
) and cAMP ꢀLog(τ/K
A
). (B) Calcium ꢀLog(τ/K
A
A
) and cAMP ꢀLog(τ/K ) values are
plotted relative to each other to visualize the relative contribution of each value to the calculated
Log(Bias Factor). Values that fall upon the slope of unity (black dotted line) represent ligands with
no bias between calcium mobilization and cAMP inhibition. Agonists that fall upon the slope of 1
(red dotted line) have a similar bias as that of 1, as there is a similar magnitude of difference
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
between the calcium ꢀLog(τ/K ) and cAMP ꢀLog(τ/K ) values. Data represent the mean ± SEM
A
A
from 3 to 12 independent experiments performed in duplicate or triplicate.
Acetylation or removal of the amino group on the allosteric thiophene pharmacophore (c.f.
compound 9 and 10, respectively) did not significantly influence the bias profile when compared to
the parent compound; although 9 trended towards an increase in bias relative to 1 (Figure 2A).
Similarly, the removal of the benzoyl group (compound 11) had no significant effect on the bias
profile observed (Figure 2A). However, removal of the 4ꢀ(trifluoromethyl)phenyl group
(compound 12) or two or more of the thiophene substituents (compounds 13–16) significantly
reduced the bias away from calcium mobilization, relative to compound 1 (Figure 2A).
Interestingly, compound 16, in which three of the thiophene substituents have been deleted, showed
a reversal in the bias profile relative to compound 1, such that it demonstrated significant bias
toward calcium mobilization (Figure 2A). These findings highlight the importance of the allosteric
moiety substituents, particularly the 4ꢀ(trifluoromethyl)phenyl on the thiophene scaffold in retaining
a biased signaling profile away from calcium mobilization.
Investigating the impact of structural modifications to adenosine receptor subtype selectivity.
To investigate the influence of structural modifications on adenosine receptor subtype selectivity,
additional studies were performed using FlpInꢀCHO cells stably expressing the human A AR
2
A
3 3
(A2AARꢀFlpInꢀCHO), human A2BAR (A2BARꢀFlpInꢀCHO) or human A AR (A ARꢀFlpInꢀCHO).
At the A2AAR, all ligands except compound 8, stimulated a detectable level of cAMP accumulation,
with no significant difference in the maximal effect (Table 2). A significant increase in the A2AAR
pEC50 and Log(τ/K
A
), relative to 1, was observed upon modification of the 2ꢀ and 5’ꢀpositions of
ACS Paragon Plus Environment

Page 15 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
orthosteric pharmacophore (compounds 3 and 4, respectively), relocation of the thiophene ring from
the para to the metaꢀposition (compound 7), and deletion of the trifluoromethylphenyl substituent
of the thiophene ring (compound 12) (Table 2). At the A2BAR, the structural modifications had no
significant effect on the pEC , E
or Log(τ/K ), with the exception of 16, which had a
50
MAX
A
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
significantly increased Log(τ/K
activity at any concentration employed (up to 10 ꢁM). Additionally, the activity of 5 and 10 was
too low to be adequately defined. An increase in the Log(τ/K ), relative to 1, was observed for 6, 7
and 8, which have modified linkers, and 15, which has two substituents deleted from the thiophene
Table 2). Relative to NECA, the parent compound 1, and the derivatives 6, 9, 11 and 13 were
A 3
) (Table 2). At the A AR, compounds 2 and 3 had no significant
A
(
partial agonists at the A AR (Table 2).
3
1
The influence of structural modifications on the selectivity index, relative to the A AR, was
established at each adenosine receptor subtype for modulation of cAMP accumulation. Agonist
Log(τ/K ) values were normalized to the corresponding NECA Log(τ/K ), thereby forming
Log(τ/K ). This value represents a system independent measurement of agonist activity at each
adenosine receptor subtype, which was subsequently normalized relative to the A AR to establish
the selectivity index and thus quantify subtype selectivity. The selectivity index reflects a ratio of
the ꢀLog(τ/K ) determined for each agonist at each subtype, relative to the ꢀLog(τ/K ) determined
for the A AR. The selectivity index is expressed on a log scale, whereby a selectivity index equal
A
A
ꢀ
A
1
A
A
1
to zero reflects no subtype selectivity between the A
NECA. In contrast, values greater than or less than zero reflect an increase or decrease in subtype
selectivity relative to the A AR, respectively.
1
AR and corresponding subtype, relative to
1
The selectivity index, normalized to the nonꢀselective agonist NECA, demonstrated that 1 had
greatest selectivity for the A AR, followed by the A AR, A AR then A AR. Modification of the
2B
1
3
2A
orthosteric pharmacophore (compounds 2ꢀ4) resulted in a significant increase in the A2AAR
selectivity index, with each of these compounds demonstrating preference for the A2AAR over the
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
A
1
AR (Figure 3). Similarly, 12ꢀ14 and 16, which incorporated thiophene substituent deletions,
resulted in an increase in the A AR selectivity index, such that these compounds were nonꢀ
2
A
selective for the A
by one carbon, exhibits significantly greater A
metaꢀsubstituted) had modest A2AAR selectivity, whereas 8 (orthoꢀsubstituted) showed no
appreciable A2AAR activity, and therefore exhibited A AR selectivity (Figure 3). As observed
previously, 1 is a relatively high potency A AR agonist, with greater selectivity relative to
1
AR and A2AAR (Figure 3). Compound 6, in which the linker length is increased
1
AR selectivity relative to the A2AAR. Compound 7
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
1
3
1
2B
NECA, for the A2BAR over the A
degree of selectivity to that of 1, some modifications had significantly greater preference for the
2BAR. These included derivatives with modification to the 2ꢀposition of the purine group (2 and
) and complete truncation of substituents on the thiophene ring (16), each of which had an
approximate 200ꢀfold greater A AR selectivity index over the other adenosine receptor subtypes.
1
AR (Figure 3). Whilst most derivatives maintained a similar
A
3
2B
The increase in A2BAR subtype selectivity relative to the A
1
AR, for compounds 2 and 3 resulted
predominantly from a decrease in the A AR Log(τ/K ), whereas the increased A AR subtype
1
A
2B
selectivity relative to the A
AR Log(τ/K ), respectively. A significant increase in the A
when the allosteric pharmacophore was shifted from the para to the meta or orthoꢀposition
compounds 7 and 8, respectively). Indeed, 7 and 8 had approximately 10ꢀfold and 100ꢀfold greater
relative selectivity for the A AR than the A AR, respectively (Figure 3). These findings suggest
1
AR, for 16 is due to combined increase and decrease in the A2BAR and
A
1
A
3
AR selectivity index was observed
(
3
1
modification of the linker region may enable the generation of a bivalent ligand with A
selectivity.
3
AR
ACS Paragon Plus Environment

Page 17 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 3. The subtype selectivity of 1 and derivatives at the four adenosine receptor subtypes,
relative to the nonꢀselective agonist NECA. The subtype selectivity index was utilized to attain a
systemꢀindependent measurement of agonist subtype selectivity. Values are logarithmic and, when
equal to zero, indicate no subtype selectivity relative to the A AR, whereas values greater than or
1
1
less than zero reflect an increase or decrease in the subtype selectivity relative to the A AR,
respectively. *P<0.05, oneꢀway ANOVA Dunnett’s postꢀhoc analysis when compared to the
#
selectivity index of 1 at the corresponding adenosine receptor subtype. Subtype selectivity index
could not be determined due to minimal activity. Data represent the mean ± SEM from 3 to 10
independent experiments performed in duplicate or triplicate.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
Table 2. Subtype Selectivity of VCP746 and Derivatives in Human AR Expressing FlpIn-CHO Cells at the cAMP Pathway.
2
NH
Orthosteric Pharmacophore
2-NH
3 2-Cl
Allosteric Pharmacophore
2-NHC(O)CH
10 des 2-NH2
S
2
2
2
9
3
H
3
HN (CH2)n
N
N
O
4
4 5'-C(O)NHEt
11 des 3-C(O)Ph
12 des 4-PhCF3
13 des 2-NH2 + des 3-C(O)Ph
1
15 des 3-C(O)Ph + des 4-PhCF3
16 des 2-NH2 + des 3-C(O)Ph + des 4-PhCF3
N
N
O
Linker Modifications
5 n= 5
6
7
2
HO 5'
3
F C
N
4 des 2-NH2 + des 4-PhCF3
O
n= 7
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
meta thiophene
HO
OH
8 ortho thiophene
Cpd
NECA
A
1
-FlpInCHO
A
2A-FlpInCHO
A
2B-FlpInCHO
7.80 0.18
40 ± 3
A
3
-FlpInCHO
7.85 0.09
97 ± 3
pEC50
9.12 0.08
±
8.48 0.10
±
±
±
E
Max
101 ± 3
8.98 ± 0.09
10
105 ± 3
8.53 ± 0.08
6
Log(τ/KA)
n
7.65 ± 0.16
7
7.94 ± 0.08
6
1
2
3
4
5
pEC50
7.37
±
0.07
5.64
±
0.11
7.02
37 ± 5
6.80 ± 0.22
4
7.01
43 ± 4
7.03 ± 0.18
4
7.73
39 ± 3
7.60 ± 0.19
4
7.12
38 ± 4
6.97 ± 0.18
4
±
0.30
6.10 ± 0.17
49 ± 6
5.94 ± 0.17
4
ND
ND
ND
3
###
E
Max
112 ± 4
7.35 ± 0.07
3
5.87
115 ± 4
6.01 ± 0.04****
4
6.65
106 ± 5
6.70 ± 0.03**
4
7.01
112 ± 8
6.80 ± 0.09**
4
7.02
114 ± 6
6.98 ± 0.08
3
109 ± 9
5.67 ± 0.09
3
Log(τ/KA)
n
pEC50
EMax
±
0.05****
6.12
±
0.16
±
0.22
111± 11
6.07 ± 0.12
3
6.84
110 ± 5
6.78 ± 0.13****
3
7.25
114 ± 6
7.15 ± 0.12****
3
5.44
131 ± 34
5.56 ± 0.09
3
Log(τ/K
n
pEC50
A
A
A
)
)
)
±
0.11**
±
0.10***
±
0.23
ND
ND
ND
3
E
Max
Log(τ/K
n
pEC50
±
0.16
±
0.12****
±
0.24
5.85
± 0.17
E
Max
75 ± 8
5.51 ± 0.13
4
<6
ND
<6
3
Log(τ/K
n
pEC50
±
0.08
±
0.27
7.20
52 ± 5
7.58 ± 0.23
4
±
0.22
E
Max
Log(τ/KA)
n
ACS Paragon Plus Environment

Page 19 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
6
7
8
9
pEC50
7.64
±
0.10
5.31
±
0.13
6.33
38 ± 5
6.18 ± 0.17
5
7.45
43 ± 5
7.45 ± 0.26
3
6.54
46 ± 5
6.66 ± 0.19
4
6.21
44 ± 8
6.25 ± 0.23
4
6.83
49 ± 5
7.03 ± 0.22
4
6.35
28 ± 6
6.15 ± 0.34
4
6.65
40 ± 6
6.56 ± 0.27
3
7.04
44 ± 5
7.10 ± 0.25
3
±
0.26
6.48
±
0#.29
E
Max
105 ± 5
7.85 ± 0.13*
3
6.98
99 ± 9
6.97 ± 0.23
3
5.89
104 ± 15
5.89 ± 0.15****
3
97 ± 13
5.20 ± 0.10*
3
6.83
83 ± 10
6.75 ± 0.13****
3
ND
ND
ND
3
56 ± 13
6.64 ± 0.23*
3
6.91
80 ± 6
6.90 ± 0.09**
3
6.55 ± 0.21
103 ± 11
6.80 ± 0.09**
3
6.35
60 ± 25
6.36 ± 0.23
3
<6
ND
<6
3
Log(τ/K
n
pEC50
A
)
±
0.19
±
0.21***
±
0.30
± 0.19
E
Max
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
Log(τ/KA)
n
pEC50
±
0.20****
±
0.22
E
Max
Log(τ/K
n
pEC50
EMax
Log(τ/K
n
A
)
A
)
A
)
A
)
6.86
±
0.11
5.63
±
0.37
±
0.31
±
0#.47
119 ± 9
6.97 ± 0.06
3
91 ± 31
5.35 ± 0.09
3
1
0
1
2
pEC50
6.96
±
0.08
5.54
±
0.14
±
0.24
E
Max
114 ± 6
6.93 ± 0.11
3
101 ± 13
5.48 ± 0.09
3
Log(τ/K
n
pEC50
1
1
7.42
±
0.13
5.57
±
0.11
±
0.40
6.55
58 ± 6
6.23 ± 0.16
3
±#0.17
E
Max
110 ± 6
7.44 ± 0.10
3
127 ± 12
5.72 ± 0.09
3
6.45
110 ± 5
6.69 ± 0.09****
3
6.19
61 ± 9
Log(τ/K
n
pEC50
7.38
±
0.11
±
0.08*
±
0.35
6.35 ± 0.21
E
Max
101 ± 5
7.21 ± 0.04
3
6.77
106 ± 5
85 ± 11
6.22 ± 0.09
3
6.46 ± 0.27
26 ± 5
Log(τ/KA)
n
pEC50
13
±
0.08*
±
0.22
±
0.29
#
###
E
Max
Log(τ/K
n
A
)
6.64 ± 0.15***
3
6.03 ± 0.19
3
5.68 ± 0.37
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
14
pEC50
6.86
±
0.17
6.36
86 ± 7
6.10 ± 0.09
3
±
0.15
6.80
27 ± 5
6.59 ± 0.36
3
7.22
45 ± 4
7.33 ± 0.19
4
7.75
42 ± 3
7.74 ± 0.19*
4
±
0.40
6.39
74 ± 7
6.06 ± 0.10
3
7.19
72 ± 5
6.76 ± 0.09*
3
± 0.15
E
Max
109 ± 9
6.86 ± 0.09*
3
Log(τ/K
n
pEC50
A
)
15
7.17
±
0.17
5.97
±
0.12
±
0.20
± 0.15**
E
Max
100 ± 7
7.04 ± 0.15
3
6.63
117 ± 6
106 ± 8
6.02 ± 0.09
3
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
Log(τ/KA)
n
pEC50
16
±
0.10**
5.88
±
0.10
±
0.21
6.08 ± 0.10
E
Max
105 ± 9
6.11 ± 0.10
3
107 ± 7
6.24 ± 0.09
3
Log(τ/K
n
A
)
6.69 ± 0.09**
3
Data represents the mean ± SEM of 3 to 10 separate experiments conducted in duplicate or triplicate (A
clarity). * P<0.05, ** P<0.01, *** P<0.001, **** P<0.0001 derivative value was significantly different (oneꢀway analysis of variance, Dunnett’s postꢀ
hoc) compared to the corresponding parameter for 1 at the corresponding adenosine receptor subtype. # P<0.05, P<0.01,
different (oneꢀway analysis of variance, Dunnett’s postꢀhoc) when compared to the NECA EMAX at the corresponding adenosine receptor subtype.
1
ꢀFlpInCHO data from Table 1 included for
##
###
P<0.001, significantly
ACS Paragon Plus Environment

Page 21 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
Molecular Modeling. Molecular modelling employed an active A AR homology model based
on the recent crystal structure of A AR coꢀbound with agonist NECA and an engineered G
2
A
32
protein. Incorporated within the active A
1
AR homology model were several extracellular loop
30
(
ECL) features present in the A
1
AR inactive crystal structure (PDB ID: 5UEN) such as the
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
disulfide bonds, the ECL2 alpha helices and the ECL1/ECL2 short beta sheet. Adenosine docked
within the transmembrane (TM) bundle, where the adenine ring was stabilized by a hydrogen bond
6
.55
ECL2
with the conserved residue, N254
The orientation of adenine moiety within this active A
reported for the agonistꢀbound A2AAR crystal structures (PDB ID: 2YDO; PDB ID: 5G53).
Docking of compound 1 within the active A AR homology model, used a template docking method
and πꢀstacking interaction with ECL2, F171
(Figure 4).
1
AR homology model was similar to that
3
2,33
1
where the adenosine moiety was constrained to the docking pose determined for adenosine alone.
The linker region of compound 1 was stabilized by a hydrogen bond between the amide group and
ECL2
ECL2
E170
and hydrophobic interaction with E172
(Figure 4). The allosteric moiety of
compound 1 was predicted to bind to an extracellular cage bordered by the top of TM5, TM6, the
alphaꢀhelical region of ECL2 and ECL3. Specifically, the hydrophilic surface of this moiety was
oriented toward ECL3, with the 2ꢀamino and the 3ꢀketo groups of the thiophene ring anchored by
ECL3
hydrogen bond interactions with S262
, whereas the 4ꢀ(trifluoromethyl)phenyl group of the
ECL2
ECL2
6.58
ECL3
thiophene ring formed hydrophobic interaction with E172
, K173
, T257
and H264
(
Figure 4).
Deletion of the 4ꢀ(trifluoromethyl)phenyl group resulted in a loss of biased agonism suggesting
that predicted hydrophobic interactions within the extracellular vestibule are key for A AR biased
agonism (Figure 2, Figure 4). This finding is in accordance with previous studies demonstrating
that the addition of the 4ꢀ(trifluoromethyl)phenyl moiety can enhance the A AR allosteric
1
1
modulator activity of 2ꢀaminoꢀ3ꢀbenzoylthiopthenes (2A3BTs). Interestingly, certain modifications
of the 2ꢀaminoꢀ3ꢀbenzoylthiopthene (2A3BT) unit that are known to have a significant effect on
3
4
allosteric activity (e.g. acetylation or removal of the 2ꢀamino group in compounds 9 and 10) have
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
afforded bitopic ligands with a biased profile similar to 1 (Figure 2). This may well be a result of
the 2A3BT moiety of these bitopic compounds adopting a different binding pose relative to standꢀ
alone 2A3BT allosteric modulators. Indeed, the predicted binding pose for the allosteric
pharmacophore of compound 1, is different to the predicted binding pose of the corresponding
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
29
standꢀalone 2A3BT. It is unsurprising that the attachment of a very large linker and adenosine
moiety to the relatively low affinity allosteric modulator might alter, to some extent, the binding
mode of the allosteric moiety within compound 1.
1
Figure 4. Docking of compound 1 (green) and adenosine (cyan) into an active A AR
homology model based on an active A2AAR crystal structure (PDB ID: 5G53). Residues
contributing to hydrogen bonds and πꢀstacking interactions are highlighted in magenta,
whereas residues that form key hydrophobic interactions are highlighted in yellow.
ACS Paragon Plus Environment

Page 23 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
ꢀ
CONCLUSIONS
This study has provided the first insight into the molecular determinants and SAR of the novel
bitopic ligands at the A AR. Overall, our findings suggest that the allosteric pharmacophore,
particularly the 4ꢀ(trifluoromethyl)phenyl substituent, has an important role in governing bitopic
ligand bias at the A AR. Importantly, single deletion of the 4ꢀ(trifluoromethyl)phenyl substituent
1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
(12) abolished the bias whereas removal of all three substituents of the thiophene scaffold (16)
resulted in a reversal in the bias, such that at the A AR, compound 16 was significantly biased
1
towards calcium mobilization. Furthermore, these data highlight that extending the alkyl linker
length by one carbon (n = 7, compound 6) trended to increase both the bias and A
relative to 1. As such, compounds 6, 12 and 16 represent novel tools to interrogated the
mechanistic basis of A AR bias in cardiac (patho)physiology. Our docking studies suggest that
1
AR selectivity,
1
interactions of the allosteric moiety and linker region within the extracellular vestibule may underlie
the initial trigger of the biased agonist profile of compound 1. Interestingly, the predicted binding
6
pose for tethered allosteric moiety within the N ꢀsubstituted adenosine derivative 1 is different to
2
9
the corresponding untethered allosteric modulator. Collectively, these findings will inform future
structureꢀbased drug design efforts for the development of clinically relevant adenosine receptor
biased agonists.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ꢀ
EXPERIMENTAL SECTION
Chemistry. Chemicals and solvents were purchased from standard suppliers and used without
®
further purification. Davisil silica gel (40ꢀ63 µm) for flash column chromatography was supplied
by Grace Davison Discovery Sciences (Victoria, Australia) and deuterated solvents were purchased
from Cambridge Isotope Laboratories, Inc. (USA, distributed by Novachem PTY. Ltd, Victoria,
Australia).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Unless otherwise stated, reactions were carried out at ambient temperature. Reactions were
monitored by thin layer chromatography on commercially available preꢀcoated aluminiumꢀbacked
plates (Merck Kieselgel 60 F254). Visualisation was by examination under UV light (254 and 366
nm). All organic extracts collected after aqueous workꢀup procedures were dried over anhydrous
4 2 4
MgSO or Na SO before gravity filtering and evaporation to dryness. Organic solvents were
evaporated in vacuo at ≤ 40°C (water bath temperature). Purification using preparative layer
chromatography (PLC) was carried out on Analtech preparative TLC plates (200 mm x 200 mm x 2
mm).
1
13
H NMR and C NMR spectra were recorded on a Bruker Avance Nanobay III 400MHz
Ultrashield Plus spectrometer at 400.13 MHz and 100.62 MHz, respectively. Chemical shifts (δ)
are recorded in parts per million (ppm) with reference to the chemical shift of the deuterated
solvent. Coupling constants (J) and carbonꢀfluorine coupling constants (JCF) are recorded in Hz
and the significant multiplicities described by singlet (s), doublet (d), triplet (t), quadruplet (q),
broad (br), multiplet (m), doublet of doublets (dd), doublet of triplets (dt).
Preparative HPLC was performed using an Agilent 1260 infinity coupled with a binary
preparative pump and Agilent 1260 FCꢀPS fraction collector, using Agilent OpenLAB CDS
software (Rev C.01.04), and an Altima 5µM C8 22 x 250 mm column. The following buffers were
ACS Paragon Plus Environment

Page 25 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
used; buffer A: H
2
O; buffer B: MeCN, with sample being run at a gradient of 30% buffer B to
100% buffer B over 10 min, at a flow rate of 20 mL/min.
LCꢀMS were run to verify reaction outcome and purity using an Agilent 1200 Series HPLC
coupled to an Agilent 6100 Series Single Quad MS system or an Agilent 1260 Series HPLC
coupled to an Agilent 6120 Series Single Quad MS system. The following buffers were used;
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
buffer A: 0.1% formic acid in H O; buffer B: 0.1% formic acid in MeCN. The following gradient
was used with a Poroshell 120 ECꢀC18 50 x 3.0 mm 2.7 micron column, and a flow rate of 0.5
mL/min and total run time of 5 min; 0–1 min 95% buffer A and 5% buffer B, from 1ꢀ2.5 min up to
0% buffer A and 100% buffer B, held at this composition until 3.8 min, 3.8–4 min 95% buffer A
and 5% buffer B, held until 5 min at this composition. All screening compounds were of > 95%
purity. Mass spectra were acquired in positive and negative ion mode with a scan range of 100–
1000 m/z. UV detection was carried out at 214 and 254 nm. All retention times (t ) are quoted in
R
min. HRMS analyses were carried out on an Agilent 6224 TOF LC/MS Mass Spectrometer
coupled to an Agilent 1290 Infinity (Agilent, Palo Alto, CA). All data were acquired and reference
mass corrected via a dualꢀspray electrospray ionisation (ESI) source. Acquisition was performed
using the Agilent Mass Hunter Data Acquisition software version B.05.00 Build 5.0.5042.2 and
analysis was performed using Mass Hunter Qualitative Analysis version B.05.00 Build 5.0.519.13.
General amide coupling procedure used for the synthesis of final compounds 1-16.
Respective acid (1.0 eq) and respective amine (1.0 eq) were dissolved in anhydrous DMF (3 mL/
3
100 mg), to this solution was added BOP (1.5 eq) followed by Et N (5.0 eq). The reaction was
stirred at rt until full conversion was observed, followed by an individual workꢀup and purification
procedure.
4-(5-Amino-4-benzoyl-3-(3-(trifluoromethyl)phenyl)thiophen-2-yl)-N-(6-((2-amino-9-
((2R,3R,4S,5R)-3,4-dihydroxy-5-(hydroxymethyl)tetrahydrofuran-2-yl)-9 -purin-6-
H
2
yl)amino)hexyl)benzamide (2). EtOAc (100 mL) was added before being washed with H O (2 ×
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
5
2 4
0 mL) and brine (1 × 50 mL). The organic layer was dried with Na SO , filtered and the solvent
removed under reduced pressure. Purification by column chromatography (eluent DCM: MeOH
1
1
8
7
5
1
00:0 → 90:10) gave the titled product as a yellow solid (76 mg, 43%). H NMR (d
6
ꢀDMSO) δ
.39 (t, J = 5.6 Hz, 1H), 8.27 (s, 2H), 7.95 (s, 1H), 7.72–7.65 (m, 2H), 7.39–7.26 (m, 2H), 7.23–
.12 (m, 6H), 7.08–6.98 (m, 4H), 5.85–5.74 (m, 3H), 5.58–5.48 (m, 1H), 5.40 (d, J = 6.3 Hz, 1H),
.15 (d, J = 4.5 Hz, 1H), 4.56 (dd, J = 11.4, 6.2 Hz, 1H), 4.18–4.10 (m, 1H), 3.96 (q, J = 3.4 Hz,
H), 3.74–3.64 (m, 1H), 3.63–3.54 (m, 1H), 3.48–3.41 (m, 2H), 3.31–3.20 (m, 2H), 1.67–1.48 (m,
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
3
4H), 1.43–1.30 (m, 4H); C NMR (d
6
ꢀDMSO) δ 192.2, 166.4, 165.8, 160.4, 155.5, 140.6, 137.3,
36.5, 136.3, 135.2, 134.7, 133.1, 130.3, 129.2, 128.9, 128.8 (q, J = 31.9 Hz), 128.4, 127.7, 127.5,
27.4 (q, J = 3.7 Hz), 124.2 (q, J = 272 Hz), 123.6 (q, J = 3.9 Hz), 119.2, 115.9, 87.5, 86.0, 73.7,
1
1
7
–
f
1.2, 62.2, 39.9, 39.6, 29.6 (2C), 26.8, 26.7; LCMS R (min) = 3.19. MS m/z 831.3 (M + H); HRMS
+
C H F N O S [M+H] calcd 831.2900, found 831.2927.
41
42
3
8
6
4
-(5-Amino-4-benzoyl-3-(3-(trifluoromethyl)phenyl)thiophen-2-yl)- -(6-((2-chloro-9-
N
((2
R
,3
R
,4 ,5 )-3,4-dihydroxy-5-(hydroxymethyl)tetrahydrofuran-2-yl)-9H-purin-6-
S
R
yl)amino)hexyl)benzamide (3). EtOAc (100 mL) was added before being washed with H
0 mL) and brine (1 × 50 mL). The organic layer was dried with Na SO , filtered and the solvent
removed under reduced pressure. The crude product was purified twice by column chromatography
2
O (2 ×
5
2
4
(
eluent DCM: MeOH 100:0 → 90:10) to give the titled product as a yellow solid (8 mg, 6%).
1
H NMR (d
.01–6.94 (m, 2H), 5.92 (d, J = 6.2 Hz, 1H), 4.71–4.66 (m, 1H), 4.33 (dd, J = 5.1, 2.9 Hz, 1H), 4.17
(q, J = 2.7 Hz, 1H), 3.91 (dd, J = 12.5, 2.6 Hz, 1H), 3.77 (dd, J = 12.5, 2.9 Hz, 1H), 3.56 (br t, J =
4
ꢀMeOD) δ 8.22 (s, 1H), 7.64–7.58 (m, 2H), 7.20–7.13 (m, 3H), 7.12–7.03 (m, 6H),
7
6.6 Hz, 2H), 3.38–3.34 (m, 2H), 1.75–1.59 (m, 4H), 1.54–1.39 (m, 4H); LCMS R (min) = 3.41. MS
f
+
m/z 849.7 (M + H); HRMS – C41
H40ClF
3
N
7
O
6
S [M+H] calcd 850.2401, found 850.2399.
(2S,3S,4R,5R)-5-(6-((6-Aminohexyl)amino)-9H-purin-9-yl)-N-ethyl-3,4-
dihydroxytetrahydrofuran-2-carboxamide (4). The reaction mixture was evaporated to dryness
ACS Paragon Plus Environment

Page 27 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2
and the residue was taken up in EtOAc (200 mL) before being washed with H O (3 × 100 mL).
The organic layer was dried with Na SO , filtered and evaporated to dryness. Purification by
2
4
column chromotography (eluent DCM: MeOH 9:1) gave the titled compound as a yellow oil (24
1
mg, 15%). H NMR (d
3
ꢀMeOD) δ 8.27 (s, 1H), 8.24 (s, 1H), 7.62–7.59 (m, 2H), 7.18–7.14 (m,
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
H), 7.11–7.04 (m, 6H), 6.99–6.95 (m, 2H), 6.02 (d, J = 7.72 Hz, 1H), 4.76 (dd, J = 4.8 Hz, 1H),
.48 (d, J = 1.52 Hz, 1H), 4.33 (dd, J = 1.28 Hz, 1H), 3.59 (br s, 2H), 3.40–3.32 (m, 4H), 1.72
4
(quin, J = 7.08 Hz, 2H), 1.62 (quin, J = 7.36 Hz, 2H), 1.48–1.42 (m, 4H), 1.21 (t, J = 7.28 Hz, 3H);
1
3
3
C NMR (d ꢀMeOD) δ 193.5, 170.7, 168.0, 167.6, 140.4, 137.2, 137.0, 135.2, 134.1, 132.6, 129.8,
1
1
2
29.5, 129.35 (q, J = 30.3 Hz), 128.8, 128.7, 128.3, 127.9 (2C), 127.3 (q, J = 4.1 Hz), 127.2, 126.9,
23.8 (q, J = 271.5 Hz), 123.0 (q, J = 3.8 Hz), 89.1, 85.1, 73.6, 72.0, 119.8, 115.7, 53.4, 39.5, 33.7,
f
9.0 (2C), 26.4, 26.2, 13.8; LCMS R (min) = 3.41. MS m/z 857.3 (M + H); HRMS –
+
C H F N O S [M+H] calcd 857.3057, found 857.3080.
43
44
3
8
6
4
-(5-Amino-4-benzoyl-3-(3-(trifluoromethyl)phenyl)thiophen-2-yl)- -(5-((9-
N
((2R
,3 ,4 ,5 )-3,4-dihydroxy-5-(hydroxymethyl)tetrahydrofuran-2-yl)-9H-purin-6-
R
S
R
yl)amino)pentyl)benzamide (5). The product mixture was added dropwise via syringe to 100 mL
ice/water, a bright yellow precipitate formed, the suspension was stirred for 30 min and was
o
allowed to settle in the fridge (ꢀ5 C). The solution was filtered and the solid washed with cold
H
2
O, before being dried under vacuum for 12 h. This gave 196 mg crude solid which was dry
loaded onto a silica column (eluent CHCl : MeOH: NH OHaq 90:10:1). Concentration of the
appropriate fractions gave a residue, which was triturated with cold Et
3
4
2
O. Filtration and washing
1
with cold Et O, before drying under vacuum gave a yellow solid (156 mg, 90%). H NMR (d6ꢀ
2
DMSO) δ 8.35–8.31 (m, 2H), 8.23–8.17 (m, 3H), 7.90–7.84 (m, 1H), 7.65–7.60 (m, 2H), 7.27–7.22
(
(
(
m, 1H), 7.15–7.08 (m, 6H), 7.00–6.94 (m, 4H), 5.87 (d, J = 6.0 Hz, 1H), 5.45–5.42 (m, 2H), 5.17
d, J = 4.4 Hz, 1H), 4.61 (q, J = 6.0 Hz, 1H), 4.16–4.12 (m, 1H), 3.96 (q, J = 3.2 Hz, 1H), 1.60
1
3
6
quin, J = 7.2 Hz, 2H), 1.51 (quin, J = 7.2 Hz, 2H), 1.36–1.28 (m, 2H); C NMR (d ꢀDMSO) δ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
91.7, 165.9, 165.3, 154.6, 152.3, 148.2, 140.2, 139.6, 136.8, 136.1, 134.8, 134.2, 132.6, 129.8,
128.7, 128.5, 128.3 (q, J = 31.3 Hz), 127.9, 127.2 (2C), 126.9 (q, J = 3.9 Hz), 123.8 (q, J = 270.
Hz), 123.2 (q, J = 4.2 Hz), 119.9, 118.8, 115.4, 88.0, 85.9, 73.4, 70.7, 61.7, 39.2, 28.8, 23.9; HRMS
+
–
C
40
H
39
3
F N
7
O
6
S [M+H] calcd 802.2629, found 802.2626.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4
-(5-Amino-4-benzoyl-3-(3-(trifluoromethyl)phenyl)thiophen-2-yl)- -(7-((9-
N
((2
R
,3 ,4 ,5 )-3,4-dihydroxy-5-(hydroxymethyl)tetrahydrofuran-2-yl)-9H-purin-6-
R
S
R
yl)amino)heptyl)benzamide (6). The product mixture was added dropwise via syringe to 100 mL
ice/water, a bright yellow precipitate formed, the suspension was stirred for 30 min and was
o
allowed to settle in the fridge (ꢀ5 C). The solution was filtered and the solid washed with cold
H
2
O, before being dried under vacuum for 12 h. This gave 195 mg crude solid which was dry
loaded onto a silica column (eluent CHCl : MeOH: NH OHaq 90:10:1). Concentration of the
3
4
appropriate fractions gave a residue, which was triturated with cold Et O. Filtration and washing
2
1
with cold Et
2
O, before drying under vacuum gave a yellow solid (127 mg, 72%). H NMR (d
6
ꢀ
DMSO) δ 8.34–8.30 (m, 2H), 8.22–8.19 (m, 3H), 7.91–7.86 (m, 1H), 7.62 (dt, J = 6.8, 2.0 Hz, 2H),
7.26–7.23 (m, 1H), 7.14–7.07 (m, 6H), 7.00–6.94 (m, 4H), 5.87 (d, J = 6.2 Hz, 1H), 5.44–5.42 (m,
2
1
H), 5.17 (d, J = 4.4 Hz, 1H), 4.61 (q, J = 6.2 Hz, 1H), 4.16–4.12 (m, 1H), 3.96 (q, J = 3.2 Hz, 1H),
13
.56 (quin, J = 6.4 Hz, 2H), 1.45 (quin, J = 6.4 Hz, 2H), 1.34–1.22 (m, 6H); C NMR (d ꢀDMSO)
6
δ 191.7, 165.9, 165.3, 154.7, 152.3, 148.2, 140.2, 139.6, 136.8, 136.1, 134.8, 134.2, 132.6, 129.9,
28.7, 128.4, 128.4 (q, J = 31.4 Hz), 127.9, 127.2 (2C), 127.0 (q, J = 4.2 Hz), 123.8 (q, J = 270.9
1
Hz), 123.2 (q, J = 3.6 Hz), 119.8, 118.8, 115.4, 88.0, 85.9, 79.2, 73.5, 70.7, 61.7, 39.2, 29.0, 28.6,
+
26.5, 26.4; HRMS – C H F N O S [M+H] calcd 830.2942, found 830.2937.
42
43
3
7
6
3
-(5-Amino-4-benzoyl-3-(3-(trifluoromethyl)phenyl)thiophen-2-yl)-
,3 ,4 ,5 )-3,4-dihydroxy-5-(hydroxymethyl)tetrahydrofuran-2-yl)-9H-purin-6-
yl)amino)hexyl)benzamide (7). The mixture was slowly diluted with H O (50 mL) and extracted
O (2 × 50 mL), brine
N-(6-((9-
((2R R S R
2
with EtOAc (3 × 50 mL). The combined organic layers were washed with H
2
ACS Paragon Plus Environment

Page 29 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(
50 mL) and then dried with MgSO
4
, filtered and concentrated to a yellow/orange foam (479 mg).
The crude was dissolved in a minimum of 95:5:1, CHCl : MeOH: NH OH and chromatographed
3
4
aq
3 4
on silica gel (eluent CHCl : MeOH: NH
OHaq 95:5:1 → 90:10:1). The appropriate fractions were
concentrated to a resin that was sonicated in EtOH and filtered providing an amorphous solid (54
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
mg, 9%). H NMR (d
6
ꢀDMSO) δ 8.40–8.29 (m, 2H), 8.24–8.14 (m, 3H), 7.95–7.83 (m, 1H), 7.63–
7
.54 (m, 2H), 7.23–7.16 (m, 2H), 7.16–7.04 (m, 6H), 7.00–6.91 (m, 3H), 5.88 (d, J = 6.2 Hz, 1H),
.47–5.40 (m, 2H), 5.18 (d, J = 4.5 Hz, 1H), 4.61 (dd, J = 11.4, 5.9 Hz, 1H), 4.14 (dd, J = 7.6, 4.6
5
Hz, 1H), 3.96 (dd, J = 6.1, 3.0 Hz, 1H), 3.67 (dt, J = 11.9, 3.9 Hz, 1H), 3.60–3.38 (m, 3H), 3.18 (dd,
13
J = 12.7, 6.6 Hz, 2H), 1.65–1.23 (m, 8H); C NMR (d
6
ꢀDMSO) δ 192.1, 166.2, 165.9, 155.1,
1
1
1
8
52.9, 148.7, 140.6, 140.1, 137.2, 135.4, 134.9, 134.6, 133.8, 131.9, 130.3, 129.0, 128.7, 128.5,
28.4, 128.3 (q, J = 31.2 Hz), 127.6, 127.5, 127.4, 126.1, 123.5 (q, J = 272.2 Hz), 120.2, 119.5,
15.6, 88.4, 86.4, 73.9, 71.2, 62.2, 39.0, 29.5, 29.5, 26.7, 26.7; LCMS R (min) = 5.86. MS m/z
f
+
16.2 (M + H); HRMS – C41
H
41
F
3
N
7
O
6
S [M+H] calcd 816.2786, found 816.2790.
2-(5-Amino-4-benzoyl-3-(3-(trifluoromethyl)phenyl)thiophen-2-yl)-
,3 ,4 ,5 )-3,4-dihydroxy-5-(hydroxymethyl)tetrahydrofuran-2-yl)-9
O (20 mL) and extracted
with EtOAc (3 × 50 mL). The combined organic layers were washed with H O (2 × 50 mL), brine
N
-(6-((9-
((2
R
R
S
R
H-purin-6-
yl)amino)hexyl)benzamide (8). The mixture was slowly diluted with H
2
2
(50 mL) and then dried with MgSO
4
, filtered and concentrated to a yellow gum. The crude was
: MeOH: NH
OHaq 95:5:1 → 90:10:1). The appropriate
chromatographed on silica (eluent CHCl
3
4
2
fractions were concentrated to a resin that was dissolved in EtOH and precipitated with H O (125
1
mg, 78%). H NMR (d ꢀDMSO) δ 8.34 (s, 1H), 8.20 (s, 1H), 8.11 (s, 3H), 7.86 (s, 1H), 7.36–7.24
6
(
m, 2H), 7.20–7.04 (m, 7H), 7.00–6.86 (m, 4H), 5.89 (d, J = 6.1 Hz, 1H), 5.52–5.38 (m, 2H), 5.20
d, J = 4.5 Hz, 1H), 4.62 (dd, J = 11.1, 5.7 Hz, 1H), 4.22–4.08 (m, 1H), 4.03–3.91 (m, 1H), 3.73–
(
3
.64 (m, 1H), 3.61–3.52 (m, 1H), 3.46 (s, 2H), 3.18–3.03 (m, 2H), 1.66–1.40 (m, 4H), 1.40–1.25
1
3
(m, 4H); C NMR (d
6
ꢀDMSO) δ 191.0, 168.0, 166.2, 154.7, 152.4, 148.2, 140.1, 139.7, 139.2,
ACS Paragon Plus Environment