Discovery of G Protein-Biased D2 Dopamine Receptor Partial Agonists

Article abstract of DOI:10.1021/acs.jmedchem.6b01208

Biased ligands (also known as functionally selective ligands) of G protein-coupled receptors are valuable tools for dissecting the roles of G protein-dependent and independent signaling pathways in health and disease. Biased ligands have also been increasingly pursued by the biomedical community as promising therapeutics with improved efficacy and reduced side effects compared with unbiased ligands. We previously discovered first-in-class β-arrestin-biased agonists of dopamine D2 receptor (D2R) by extensively exploring multiple regions of aripiprazole, a balanced D2R agonist. In our continuing efforts to identify biased agonists of D2R, we unexpectedly discovered a G protein-biased agonist of D2R, compound 1, which is the first G protein-biased D2R agonist from the aripiprazole scaffold. We designed and synthesized novel analogues to explore two regions of 1 and conducted structure-functional selectivity relationship (SFSR) studies. Here we report the discovery of 1, findings from our SFSR studies, and characterization of novel G protein-biased D2R agonists.

Full text of DOI:10.1021/acs.jmedchem.6b01208

Subscriber access provided by RYERSON UNIVERSITY
Article
Discovery of G Protein-biased D2 Dopamine Receptor Partial Agonists
Xin Chen, John D. McCorvy, Matthew G Fischer, Kyle V. Butler, Yudao Shen, Bryan L Roth, and Jian Jin
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b01208 • Publication Date (Web): 02 Nov 2016
Downloaded from http://pubs.acs.org on November 3, 2016
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 63
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Discovery of G Proteinꢀbiased D2 Dopamine
Receptor Partial Agonists
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
†
,#
§
,#
†
†
†
Xin Chen, John D. McCorvy, Matthew G. Fischer, Kyle V. Butler, Yudao Shen, Bryan L.
§,
†
Roth, * and Jian Jin *
†
Departments of Pharmacological Sciences and Oncological Sciences, Icahn School of Medicine
at Mount Sinai, 1425 Madison Avenue, New York, New York 10029, USA
§
Department of Pharmacology and National Institute of Mental Health Psychoactive Drug
Screening Program, School of Medicine, University of North Carolina at Chapel Hill, Chapel
Hill, North Carolina 27599, USA
RECEIVED DATE (to be automatically inserted after your manuscript is accepted)
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 63
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
ABSTRACT: Biased ligands (also known as functionally selective ligands) of G proteinꢀ
coupled receptors are valuable tools for dissecting the roles of G proteinꢀdependent and
independent signaling pathways in health and disease. Biased ligands have also been
increasingly pursued by the biomedical community as promising therapeutics with improved
efficacy and reduced side effects compared with unbiased ligands. We previously discovered
firstꢀinꢀclass βꢀarrestinꢀbiased agonists of dopamine D2 receptor (D2R) by extensively exploring
multiple regions of aripiprazole, a balanced D2R agonist. In our continuing efforts to identify
biased agonists of D2R, we unexpectedly discovered a G proteinꢀbiased agonist of D2R,
compound 1, which is the first G proteinꢀbiased D2R agonist from the aripiprazole scaffold. We
designed and synthesized novel analogs to explore two regions of 1 and conducted structure–
functional selectivity relationship (SFSR) studies. Here we report the discovery of 1, findings
from our SFSR studies, and characterization of novel G proteinꢀbiased D2R agonists.
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 3 of 63
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
INTRODUCTION
The D2 dopamine receptor (D2R) remains a main target for the development of antiꢀ
1
2
Parkinson and antipsychotic agents. Although all five dopamine receptors (D1R ꢀ D5R) are
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
3
4
expressed in brain, D2R is by far the most highly studied and is involved in many
5
6
7
neuropsychiatric diseases including attentionꢀdeficit disorders and schizophrenia. It is
therefore not surprising that all of the current FDAꢀapproved drugs for the treatment of
8
schizophrenia have direct action at D2R. Despite the prevalence of drugs targeting D2R, early
typical antipsychotic drugs are known to have debilitating motor sideꢀeffects including
9
extrapyramidal symptoms (EPS), often leading to permanent symptoms such as tardive
1
0
dyskinesia. Thus, safer and more effective drugs that target D2R are needed not only to treat
1
1
the positive and negative symptoms of schizophrenia, but also to minimize the motor sideꢀ
12
effects commonly observed with chronic antipsychotic treatment.
Newer generation
1
3
14
antipsychotics such as aripiprazole possess partial agonist actions at D2R and have less
1
5
propensity to produce motor sideꢀeffects, but mounting evidence in the last decade has
16
suggested that aripiprazole, which is a balanced D2R agonist, can activate a plethora of
1
7ꢀ18
downstream signaling pathways.
The growing realization of the complexity of G protein coupled receptor (GPCR)
mediated signal transduction pathways, specifically D2R mediated signaling pathways, has
provided a theoretical framework for the development of functionally selective or biased
4, 19
ligands,
which refer to a ligand’s ability to differentially modulate canonical (e.g., G proteinꢀ
dependent) versus nonꢀcanonical (e.g., G proteinꢀindependent, βꢀarrestinꢀmediated) signaling
20
pathways. We previously discovered novel βꢀarrestinꢀbiased agonists of D2R by extensively
1
6, 21
exploring multiple regions of the aripiprazole scaffold.
Using these βꢀarrestinꢀbiased D2R
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 63
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
agonists, which were tested in a battery of in vitro and in vivo assays, we found that the βꢀarrestin
bias appears to be important to exert superior therapeutic utility in animal models of
schizophrenia and elicit a lower level of catalepsy compared to the D2R antagonist
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
16, 22
haloperidol.
Others have recently reported the discovery of multiple functionally selective
D2R ligands based on the potential antipsychotic agent cariprazine and other D2R targeting
2
3ꢀ25,26
compounds.
In addition, the first G proteinꢀbiased agonist of D2R was recently
Biased D2R ligands are thought to stabilize different conformations of D2R leading
2
5,27
reported.
2
8ꢀ30
to the preference for either G proteinꢀdependent or G proteinꢀindependent signaling,
usually
31ꢀ34
through recruitment or inhibition of βꢀarrestin.
Future characterization of biochemical and
behavioral effects of biased D2R ligands and optimization of their drugꢀlike properties could
ultimately lead to improved antipsychotic drugs that are safer and more effective than existing
antipsychotics. While a number of βꢀarrestinꢀbiased D2R agonists have been generated and
progressed towards clinical development, there is only one reported G proteinꢀbiased D2R
27
agonist.
To develop G proteinꢀbiased D2R agonists as useful tools for elucidating the role of G
proteinꢀmediated D2R signaling in physiology and pathophysiology, we extended our structure–
functional selectivity relationship (SFSR) studies based on the aripiprazole scaffold. From these
studies, we discovered compound 1, a benzothiazole containing D2R agonist with an unexpected
bias for G protein signaling. We then explored two regions of this lead and uncovered structural
features that affect ligand bias. By combining preferred structural motifs, novel G proteinꢀbiased
D2R agonists devoid of any measurable βꢀarrestin recruitment were discovered.
RESULTS AND DISCUSSION
ACS Paragon Plus Environment

Page 5 of 63
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Discovery of Compound 1. We previously identified UNC9994 (Figure 1) as a βꢀ
2
1
arrestinꢀbiased D2R agonist that lacks Gα activity. A close analog of this compound with the
i/o
middle piperidine replaced by piperazine, UNC9995 (Figure 1), retained the βꢀarrestinꢀbiased
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
1
feature for D2R with slightly decreased potency. Adding a simple methyl group at the 2
position of the benzothiazole ring resulted in compound 2 (Figure 1), which is a balanced D2R
agonist with similar efficacy and potency in both βꢀarrestin and Gi/o pathways. Surprisingly,
extending the middle linker from propoxy (compound 2) to butoxy (compound 1, Figure 1)
resulted in the first G proteinꢀbiased D2R ligand from the aripiprazole scaffold, which exhibited
D2R Gi/o partial agonist activity (EC = 15 nM, E = 50%) with weak efficacy for βꢀarrestin
5
0
max
recruitment (EC = 13 nM, Emax = 23%).
5
0
Figure 1. Discovery of compound 1, a G proteinꢀbiased D2R partial agonist. Benzothiazole
substitution and linker modification of a βꢀarrestinꢀbiased D2R agonist led to compound 1,
which is a D2R Gi/o partial agonist with weak efficacy for βꢀarrestin recruitment.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 63
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
To develop more G proteinꢀbiased D2R agonists, we conducted structure–functional
selectivity relationship (SFSR) studies based on this newly identified G proteinꢀbiased lead. We
designed, synthesized and evaluated derivatives of compound 1 to determine which structural
modifications would favor Gi/o activation over βꢀarrestin recruitment. We previously found that
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
conformationallyꢀconstrained central linkers could lead to a significant bias for βꢀarrestin
21
recruitment over Gi/o signaling. We and others also found that the butoxy was preferred
3
5
compared to propoxy (Figure 1) or pentoxy as a central linker. Based on these findings, we
kept the butoxy constant as the central linker for the SFSR studies. We then focused our
exploration on two regions of compound 1, namely the left hand side (LHS) phenylpiperazine
and right hand side (RHS) benzothiazole moieties.
SFSR Studies of the Benzothiazole Moiety. To determine the effects of substituents at
the 2ꢀposition of the benzothiazole moiety on ligand bias, we designed the analogs of compound
1 outlined in Table 1. The synthesis of these compounds is summarized in Scheme 1. Generally,
21
our previously wellꢀestablished twoꢀstep alkylation sequence was employed to prepare these
compounds. Commercially available 2ꢀmethylbenzo[d]thiazolꢀ5ꢀol, 2ꢀmethylbenzo[d]thiazolꢀ6ꢀ
ol, 2ꢀmethylbenzo[d]oxazolꢀ5ꢀol, 2ꢀchlorobenzo[d]thiazolꢀ5ꢀol, and 5ꢀhydroxybenzo[d]thiazolꢀ
2(3H)ꢀone were reacted with 1,4ꢀdibromobutane to give the bromo intermediates, which were
then reacted with 1ꢀ(2,3ꢀdichlorophenyl)piperazine (5) to afford the target compounds 1, 3, 4, 6,
and 7. The 2ꢀtrifluoromethyl benzo[d]thiazoleꢀ5ꢀol (10) was synthesized from commercially
available 3ꢀmethoxyaniline. Trifluoroacetic acid was reacted with 3ꢀmethoxyaniline to give
.
36
intermediate 8, which was treated with NaSH H O and then cyclized to yield intermediate 9.
2
3
7
Subsequent dimethylation reaction afforded intermediate 10, which underwent the twoꢀstep
alkylation sequence to furnish compound 11. The 2ꢀethyl, 2ꢀisopropyl, 2ꢀmethylamine and 2ꢀ
ACS Paragon Plus Environment

Page 7 of 63
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
dimethylamine substituted benzothiazole building blocks were synthesized starting from
commercially available 1ꢀisothiocyanotoꢀ3ꢀmethoxybenzene. Typical Grignard reaction
procedures with ethylmagnesium chloride and isopropylmagnesium chloride yielded the
corresponding thioamide compounds 12 and 13, which were then cyclized to give intermediates
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
3
8
14 and 15. Removal of the methyl group followed by the twoꢀstep alkylation reactions afforded
compounds 18 and 19. In addition, 1ꢀisothiocyanatoꢀ3ꢀmethoxybenzene was treated with
methanamine or dimethylamine in the presence of PhCH N(CH ) Br to give
2
3 3
3
3
9
aminobenzothiazoles 21 and 22, which were subjected the deprotection and twoꢀstep alkylation
sequence to furnish the desired compounds 24 and 25.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 63
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
Scheme 1. Synthesis of compound 1 analogs to explore the RHS benzothiazole moiety.
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 9 of 63
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
Reagents and Conditions: (a) K CO EtOH, reflux, 6 h, 30ꢀ80% yield; (b) NaI/K CO , CH CN, reflux, 6 h, 50ꢀ70%
2
3,
2
3
3
o
yield; (c) (i) Ph P, Et N, CCl , TFA, 0 C, 10 min; (ii) 3ꢀmethoxyaniline/CCl , reflux, 3h, 86% yield; (d)
3
3
4
4
.
o
o
NaSH xH O, PdCl , DMSO, 50ꢀ110 C, 3 h, 85% yield; (e) HI, reflux, 5 h, 83ꢀ90%; (f) RMgBr, THF, ꢀ10 C, 90
2
2
o
o
min, ~100% yield; (g) (i) K Fe(CN) , NaOH, H O, CH OH, 60 C, 2 h, (ii) K CO , 60 C, 1 h, 82ꢀ87% yield; (h)
3
6
2
3
2
3
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
NHR’R’’, PhCH N(CH ) Br , THF, r.t., overnight, 93ꢀ96% yield.
2
3 3
3
All the newly synthesized compounds were evaluated in: (1) D2Rꢀmediated cAMP
40
accumulation assay, which measures inhibition of isoproterenolꢀstimulated cAMP production;
and (2) D2Rꢀmediated βꢀarrestinꢀ2 recruitment Tango assay to determine potency and efficacy
4
1
42
for βꢀarrestin recruitment. Quinpirole, a full agonist of D R, was used as the positive control
2
in both cAMP inhibition and βꢀarrestinꢀ2 recruitment Tango assays.
a
Table 1. SFSR of the RHS Benzothiazole Moiety.
βꢀarrestin
Emax (%)
cAMP
EC (nM)
Compound
1
RHS Group
EC (nM)
Emax (%)
50
50
23
13
50
15
3
4
6
22
25
40
76
24
48
42
36
50
29
21
43
7
59
2.6
67
5.2
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 63
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
1
1
2
2
1
8
9
4
5
40
160
N/C
N/C
N/C
N/A
32
37
140
98
<20
<20
<20
N/A
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
33
72
40
24
N/A
N/A
a
EC50 and Emax values are the average of at least 3 independent experiments performed in triplicate. Standard error
of the mean (SEM): < ± 20%. N/C. not calculated. N/A. no activity.
As summarized in Table 1, by reversing the positions of nitrogen and sulfur in the
thiazole structure, compound 3 retained similar efficacy in both βꢀarrestin and Gi/o pathways but
showed decreased potencies. The potency in βꢀarrestin recruitment decreased more than that in
Gi/o assay (5ꢀfold versus 2ꢀfold), thus favoring G protein bias. By replacing sulfur in the thiazole
ring with oxygen, the bioisostere compound 4 exhibited similar potency and efficacy in βꢀ
arrestin pathway but its efficacy in Gi/o pathway decreased, resulting in a less G proteinꢀbiased
compound compared with compound 1. With 2ꢀCl substitution, compound 6 showed similar
efficacy in Gi/o pathway compared to compound 1 whereas its efficacy in βꢀarrestin recruitment
was similar to Gi/o activity, exhibiting less G protein bias. Interestingly, the 2ꢀOH substitution,
tautomerized to 2ꢀketo, in compound 7 reversed the bias for G protein, demonstrating balanced
signaling for both Gi/o signaling and βꢀarrestin recruitment with high potencies. Incorporation of
a trifluoromethyl group, in most instances, in place of a methyl group, has been a commonly
4
3
pursued modification to lead compounds in medicinal chemistry. However, changing the
methyl group (compound 1) to trifluoromethyl group (compound 11) suffered from reduced
potency in both assays, and its efficacy in the βꢀarrestin pathway increased whereas its Gi/o
ACS Paragon Plus Environment

Page 11 of 63
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
activity decreased, making it a balanced D2R agonist like compounds 6 and 7, albeit with much
weaker potency in both pathways. Replacing the methyl group with ethyl, isopropyl, or
methylamine group (in compounds 18, 19 and 24, respectively) resulted in very weak efficacy in
βꢀarrestin recruitment (Emax < 20%) but maintained low to moderate Gi/o efficacy (Emax = 43%,
33% and 40%, respectively). Although the potency of compound 18 and 19 in the Gi/o pathway
was reduced by 5ꢀ7 fold, compound 24 was able to maintain similar potency as compound 1 in
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
Gi/o pathway (EC = 24 nM). Surprisingly, with the dimethylamine substitution, compound 25
50
did not show any activity in both pathways. Taken together, these results suggest that D2R G
protein bias is sensitive to subtle structural modifications to the 2ꢀposition of the benzothiazole
moiety and only limited substitutions are tolerated in order to achieve G proteinꢀbiased
compounds. 2ꢀMethyl, 2ꢀethyl, 2ꢀisopropyl, and 2ꢀmethylamine are possible RHS benzothizole
substitutions that may favor Gi/o signaling over βꢀarrestin recruitment.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 63
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
SFSR Studies of the Phenylpiperazine Moiety. We next investigated the LHS
phenylpiperazine moiety of compound 1. We previously reported that this moiety was very
tolerant to modifications, but most single substitutions on the phenyl ring appeared to generate
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
1
balanced compounds with similar efficacy in both βꢀarrestin and Gi/o pathways. Therefore, we
designed most compounds with 2,3ꢀdisubstitutions, including cyclized 2,3ꢀdisubstitutions on the
phenyl ring. 2ꢀMethoxy substitution on the phenyl ring was found to increase the compound’s
2
1
binding affinity towards D2R as well as the potency in both signaling pathways, therefore 1ꢀ(2ꢀ
methoxyphenyl)piperazine was selected as the only single substituted phenylpiperazine. Because
44
7
ꢀ(piperazinꢀ1ꢀyl)benzo[d]oxazolꢀ2(3H)ꢀone is present in the structure of bifeprunox, which
exhibited extraordinary potency and efficacy in both signaling pathways in our assays, we
included this moiety in our design to enhance the comparatively low efficacy and low potency of
compound 1. In addition, piperazine was replaced by piperidine, homopiperazine, or substituted
piperazine to explore whether any one of these ring replacements would favor Gi/o signaling
over βꢀarrestin recruitment. The synthesis of these compounds is outlined in Scheme 2.
a
Scheme 2. Synthesis of compound 1 analogs to explore the LHS phenylpiperazine moiety.
ACS Paragon Plus Environment

Page 13 of 63
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
O
N
S
R1
R2
R1
R2
Br
O
26
N
S
N
n
X
NH
n
X
a
R = R =Cl, X = CH, n = 1:
27
28
R = R =Cl, X = CH, n = 1:
30
31
1
2
1
2
R = R = Cl, X = N, n = 2:
R = R = Cl, X = N, n = 2:
1 2
1
2
R = H, R = OCH , X = N, n = 1: 29
R = H, R = OCH , X = N, n = 1: 32
1 2 3
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
2
3
Cl
Cl
Cl
Cl
Cl
Cl
HN
HN
N Boc
2
a
6
O
N
S
Br
N
N
N
NH
c
b
3
3
34
NH
N
N
N
2
a
6
O
N
S
d
N
O
N
Br
N
N
NH
35
36
X
HN
HN
NH
X
O
X
O
O
N
S
2
a
6
N
N
NH
d
Br
X = NH:
X = O:
39
40
X = NH:
X = O:
37
38
B(OH)2
Br
Br
N Boc
Br
N
N Boc
b
e
N
N Boc
4
1
42
26
a
O
N
S
c
N
NH
N
N
43
44
O
O
HN
O
HN
O
2
a
6
O
N
S
N
N
N
NH
4
5
a
Reagents and Conditions: (a) NaI/K CO , CH CN, reflux, 6 h, 50ꢀ70% yield; (b) Pd (dba) , BINAP, tꢀBuONa,
2
3
3
2
3
o
toluene, sealed tube, 100 C, overnight, 52ꢀ89% yield; (c) TFA, CH Cl , r.t., 1 h, 92ꢀ96% yield; (d) Pd (dba) ,
2
2
2
3
o
BINAP, tꢀBuONa, toluene, sealed tube, 100 C, 4h, 56ꢀ67% yield; (e) Pd(PPh ) , K CO , dioxane, H O, microwave,
3
4
2
3
2
9
3% yield.
21
16
4
ꢀ(2,3ꢀDichlorophenyl)piperidine (27) and 1ꢀ(2,3ꢀdichlorophenyl)ꢀ1,4ꢀdiazepane (28)
were prepared according to the previously published procedures. Compounds 30, 31, and 32
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 63
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
were prepared by the nucleophilic displacement of 5ꢀ(4ꢀbromobutoxy)ꢀ2ꢀmethylbenzo[d]thiazole
(26) with corresponding starting materials 27, 28, and commercially available 1ꢀ(2ꢀ
methoxyphenyl)piperazine (29). Buchwald–Hartwig amination between 1ꢀbromoꢀ2,3ꢀ
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
45
dichlorobenzene and Bocꢀprotected methylpiperazine, followed by the removal of the Boc
protecting group gave intermediate 33, which underwent the nucleophilic displacement reaction
with bromide 26 to furnish compound 34. Similarly, 8ꢀbromoꢀ2ꢀmethylquinoline underwent the
Buchwald–Hartwig amination reaction with unprotected piperazine yielded intermediate 35. It
was necessary to reduce the reaction time from overnight to 4 hours to minimize the overꢀreacted
byproduct. Likewise, intermediates 37 and 38 were synthesized by the reaction of piperazine
with
dihydrobenzo[b][1,4]dioxine. Nucleophilic displacement of bromide 26 with intermediates 35,
7, and 38 afforded compounds 36, 39 and 40, respectively. Overnight Buchwald–Hartwig
amination of 1,3ꢀdibromoꢀ2ꢀmethylbenzene with Bocꢀprotected piperazine yielded intermediate
1, which was reacted with phenylboronic acid under typical Suzuki coupling reaction
8ꢀbromoꢀ3,4ꢀdihydroꢀ2Hꢀbenzo[b][1,4]oxazine
and
5ꢀbromoꢀ2,3ꢀ
3
4
conditions to afford intermediate 42, Removal of the Boc protecting group, followed by the
nucleophilic displacement reaction with bromide 26, furnished compound 44. Compound 45 was
prepared from the nucleophilic displacement reaction of 7ꢀ(piperazinꢀ1ꢀyl)benzo[d]oxazolꢀ
4
4
2
(3H)ꢀone with bromide 26.
a
Table 2. SFSR of the LHS Phenylpiperazine Moiety.
ACS Paragon Plus Environment

Page 15 of 63
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
βꢀarrestin
cAMP
EC (nM)
Compound
1
LHS Group
Emax (%)
EC (nM)
Emax (%)
50
50
23
13
49
50
15
3
3
0
1
31
66
22
76
<20
N/C
1.9
32
25
8.5
61
68
3
3
4
6
N/A
N/A
N/A
N/A
<20
N/C
25
15
3
4
9
0
<20
35
N/C
3.6
29
50
5.2
0.64
4
4
N/A
N/A
N/A
N/A
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 63
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
45
88
0.36
90
0.27
a
EC50 and Emax values are the average of at least 3 independent experiments performed in triplicate. SEM: < ± 20%.
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
N/C. not calculated. N/A. no activity.
Results of these compounds in βꢀarrestinꢀ2 recruitment Tango and inhibition of cAMP
accumulation assays are summarized in Table 2. Replacing the piperazine group in compound 1
with piperidine group in compound 30 resulted in a similar bias for G protein with slightly higher
efficacy but slightly lower potency in both Gi/o and βꢀarrestin pathways. By contrast, replacing
the piperazine group (compound 1) with homopiperazine (compound 31) led to significantly
decreased efficacy in both Gi/o and βꢀarrestin signaling, resulting in a highly potent low efficacy
partial agonist at activating Gi/oꢀmediated cAMP inhibition (EC = 1.9 nM, E = 22%), which
5
0
max
was inactive for βꢀarrestin recruitment. With 1ꢀ(2ꢀmethoxyphenyl)piperazine as the LHS group,
compound 32 exhibited slight improvement in efficacy in both assays, but the potency in Gi/o
pathway dropped around 5ꢀfold compared to compound 1, resulting in less bias for G protein
signaling. Surprisingly, the methyl substitution in the middle piperazine region (compound 34)
reduced the potency and efficacy drastically in both pathways, and led to no appreciable activity
in either Gi/o signaling or βꢀarrestin recruitment. Interestingly, by replacing the 2,3ꢀ
dichlorophenyl in compound 1 with the 2ꢀmethylquinoline, compound 36 displayed considerably
reduced efficacy in both Gi/o and βꢀarrestin signaling, making compound 36 a potent G proteinꢀ
biased partial agonist with low efficacy (EC = 15 nM, E = 25%). Similarly, compound 39,
5
0
max
with the dihydrobenzooxazine as the LHS moiety, did not recruit βꢀarrestin at all, but had
improved Gi/o potency yet decreased efficacy compared to compound 1 (EC = 5.2 nM, E
=
max
5
0
2
9%). Interestingly, the close analog, compound 40, with dihydrobenzodioxine as the LHS
ACS Paragon Plus Environment

Page 17 of 63
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
moiety, displayed markedly improved potency (EC = 0.64 nM) and similar efficacy (E =
max
50
50%) for activating Gi/o signaling compared to compound 1. Although its efficacy and potency
in βꢀarrestin recruitment also improved, 40 exhibited a similar bias for G protein as 1 and 30. By
replacing the 2,3ꢀdichloro substitutions (compound 1) with the 2ꢀmethylꢀ3ꢀphenyl substitutions,
compound 44 was found to be completely inactive in both Gi/o and βꢀarrestin assays. With the
benzooxazolone as the LHS moiety, compound 45 displayed significantly improved efficacy and
potency in both Gi/o and βꢀarrestin signaling pathways, making it a potent, balanced full agonist
of D2R. In summary, although this region is generally more tolerant to modifications than the
RHS benzothiazole region, subtle structural changes can still lead to a very significant impact on
ligand bias. Our SFSR studies on this region resulted in the identification of several additional
motifs including 2ꢀmethylquinoline, dihydrobenzooxazine, and dihydrobenzodioxine that favor
Gi/o signaling over βꢀarrestin recruitment.
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
SFSR Studies of Combination Compounds. We next designed and synthesized a number
of combination compounds (outlined in Scheme 3 and Table 3), which incorporate some of the
preferred RHS and LHS structural motifs we identified from the above studies. We selected 2ꢀ
ethyl, 2ꢀisopropyl, 2ꢀmethylamino and 2ꢀdimethylaminobenzothiazolꢀ5ꢀyl and 2ꢀ
methylbenzothiazolꢀ6ꢀyl as the RHS moiety, and 4ꢀ(2,3ꢀdichlorophenyl)piperidine, 2ꢀ
methylquinoline piperazine, dihydrobenzooxazine piperazine, and dihydrobenzodioxine
piperazine as the LHS moiety. The synthetic routes for these combination compounds are
summarized in Scheme 3. These compounds (46 – 58) were prepared following the same
synthetic approach developed for compounds 1, 3 and 4 using the corresponding LHS and RHS
intermediates.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 63
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
Scheme 3. Synthesis of combination compounds.
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
Reagents and Conditions: (a) K CO EtOH, reflux, 6 h, 30ꢀ80% yield; (b) NaI/K CO , CH CN, reflux, 6 h, 50ꢀ70%
2
3,
2
3
3
yield.
We then evaluated compounds 46 – 58 in the D2R βꢀarrestinꢀ2 recruitment Tango and
Gi/oꢀmediated cAMP inhibition assays (results are summarized in Table 3). Interestingly, seven
out of the 13 combination compounds (46 – 49 and 53 – 55) were significantly biased for G
protein over βꢀarrestin recruitment with no activity in βꢀarrestin recruitment, and four other
ACS Paragon Plus Environment

Page 19 of 63
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
compounds (50, 51, 56 and 58) exhibited some extent of Gi/o bias with improved potency and/or
efficacy in the Gi/o pathway. Surprisingly, two of the combination compounds, 52 and 57, with
dihydrobenzodioxinylpiperazine as the LHS moiety and 2ꢀethyl or 2ꢀmethylamino benzothiazole
as the RHS moiety, were inactive in both assays. All seven extremely G proteinꢀbiased
compounds (46 – 49 and 53 – 55) were agonists at Gi/o signaling with low to moderate efficacy
and moderate to high potency, but did not display appreciable βꢀarrestinꢀ2 recruitment activity (<
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
20% of quinpirole). Additionally, we also tested these compounds in cells not expressing D2R to
rule out nonspecific cAMP inhibition not related to D2R, and observed no significant cAMP
inhibition at concentrations lower than 10 µM (Figure S1). Most notably, as illustrated in Figure
2, compounds 46 and 54, similar to compounds 18 and 24, were more efficacious for Gi/o
signaling with Emax 40ꢀ50% of quinpirole, and compounds 49 and 53 exhibited high Gi/o
46
potencies (EC50 < 10 nM) but lower efficacies (Emax = 30 – 40%). A bias plot comparing
relative concentration responses in the Gi/o GloSensor assay versus Tango βꢀarrestinꢀ2
recruitment activity reveals that quinpirole is equiꢀefficacious in both assays (noted by the dotted
line), but that compounds 1, 24, 46, 49, 53, and 54 deviate from quinpirole and cluster toward G
protein activation (Figure 2C). These ligands, which have unique bias profiles, are potentially
useful tools for elucidating signaling pathways that may contribute to antipsychotic efficacy
and/or sideꢀeffects.
a
Table 3. SFSR of combination compounds.
βꢀarrestin
cAMP
Cmpd
Structure
Emax (%)
EC (nM)
Emax (%)
EC (nM)
50
50
4
6
<20
N/C
43
98
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 63
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
4
4
7
8
<20
<20
N/C
N/C
33
33
72
62
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
49
<20
N/C
32
8.3
5
5
5
5
5
0
1
2
3
4
33
36
6.4
16
41
48
1.0
2.8
N/A
5.5
50
N/A
<20
<20
N/A
N/C
N/C
N/A
37
47
5
5
5
6
<20
N/C
31
28
29
N/A
24
6.5
N/A
460
49
N/A
56
1.1
N/A
88
57
5
8
a
EC50 and Emax values are the average of at least three independent experiments performed in triplicate. SEM: < ±
0%. N/C. not calculated. N/A. no activity.
2
ACS Paragon Plus Environment

Page 21 of 63
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
From these SFSR studies, we observed the following general trends: (1) 2ꢀsubstitution at
the RHS benzothiazole moiety is a critical contributor to bias for Gi/o signaling. 2ꢀMethyl, 2ꢀ
ethyl, 2ꢀisopropyl, and 2ꢀmethylamino are preferred substituents that lead to bias for the G
protein signaling over βꢀarrestin recruitment; (2) the LHS 2,3ꢀdichlorophenyl can be replaced
with 2ꢀmethylquinoline, dihydrobenzooxazine, and dihydrobenzodioxine, resulting to significant
bias for the Gi/o signaling pathway over βꢀarrestin recruitment; (3) a small substituent such as
methyl at the middle piperazine ring can completely abolish agonist activity in both pathways
and a bulky substituent such as phenyl at the LHS ring can also completely diminish agonist
activity in both pathways; and (4) subtle ligand structural changes can result in major changes in
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
1
ligand bias, which is consistent with our previous findings.
Figure 2. Compounds 1, 24, 46, 49, 53, and 54 are G proteinꢀbiased D2R partial agonists. (A) Activity of
compounds 1, 24, 46, 49, 53, 54, and quinpirole in the D2Rꢀmediated Gi/o coupled isoproterenolꢀstimulated cAMP
production assay using HEK293T cells expressing D2R and GloSensorꢀ22F. All six tested compounds were partial
agonists compared to quinpirole (EC50 = 1.7 nM), which was used as a positive control. (B) Activity of compounds
1
, 24, 46, 49, 53, 54, and quinpirole in the D2Rꢀmediated βꢀarrestinꢀ2 translocation Tango assay using HTLA cells
transfected with a D2V2ꢀTCSꢀtTA construct. 1 recruited βꢀarrestin weakly and other five compounds did not
substantially recruit βꢀarrestin compared to the full agonist quinpirole (EC50 = 0.8 nM). Data are representative of at
least three independent experiments performed in triplicate. (C) Bias plot of compounds 1, 24, 46, 49, 53, 54, and
quinpirole in the D2Rꢀmediated cAMP inhibition GloSensor assay versus the D2Rꢀmediated βꢀarrestinꢀ2
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 63
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
translocation Tango assay. Quinpirole shows equal activity in both assays, but compounds 1, 24, 46, 49, 53, and 54
cluster toward G protein activation.
Further Characterization of G proteinꢀbiased D2R Agonists. Considering that 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
4
6
determination of ligand bias can be problematic due to either system or observational bias,
which is dependent on cell background (e.g., HEK, CHO cell lines) and signal amplification (i.e.,
47
receptor reserve, timeꢀdependence), we wanted to confirm the D2R G protein bias of these
ligands by orthologous measures of βꢀarrestinꢀ2 recruitment and G protein activity using
bioluminescence resonance energy transfer (BRET)ꢀbased assays. Aripiprazole has been
31
48
previously reported to be either a D2R antagonist or a partial agonist in the BRET βꢀarrestinꢀ
2 recruitment assay dependent on GRK2 coꢀexpression, and has been observed to act as either a
D2R antagonist or partial agonist in G proteinꢀactivation assays dependent on receptor
49
expression. Although we have quantified the extent of D2R expression comparing D2R
expressed in the GloSensor assay versus D2R expressed in the Tango assay and found similar
expression levels (Table S1), timeꢀdependent or assay kinetic differences may still interfere with
the interpretation of ligand bias.
Therefore, we chose to use the D2R βꢀarrestinꢀ2 recruitment BRET assay, where D2R
arrestin recruitment is measured after only 5 minutes, and compared compound 1 to aripiprazole
and the previously reported D2R G proteinꢀbiased ligand, MLS1547 (59)27. In HEK293T cells
coꢀexpressing D2R Cꢀterminalꢀtagged renilla luciferase (Rluc), a Venusꢀtagged βꢀarrestinꢀ2, and
G protein–coupled receptor kinase 2 (GRK2), both aripiprazole and compound 59 displayed
partial agonist activity for D2Rꢀmediated βꢀarrestinꢀ2 recruitment (Figure 3A, Emax = 26% and
6
8% of quinpirole, respectively) while compound 1 exhibited no activity up to 1 ꢀM (Figure
A). In fact, 59 also showed robust βꢀarrestin recruitment in the Tango assay (Figure 3B), albeit
3
ACS Paragon Plus Environment

Page 23 of 63
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
with moderate potency in both Tango and Gi/o assays (EC = 102 and 46 nM) (Figures 3B and
5
0
3C). To address the discrepant partial agonist activity of 59 in the βꢀarrestinꢀ2 recruitment BRET
assay compared to the previously published lack of βꢀarrestinꢀ2 recruitment as measured using
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
27
DiscoverX, we further evaluated compounds 1, 53, and 59 relative to quinpirole using the D2R
DiscoverX βꢀarrestin recruitment assay and confirmed lack of βꢀarrestinꢀ2 recruitment activity
for these compounds (Figure 3D). The lack of robust βꢀarrestinꢀ2 recruitment activity by 59
using DiscoverX assays may reflect cell type differences (DiscoverX uses CHO cells versus
BRET and Tango use HEK cell background), but also likely reflect the lack of GRK2 coꢀ
expression, where GRK2 coꢀexpression has been shown to be necessary for robust βꢀarrestinꢀ2
5
0
recruitment for the muꢀopioid receptor using the DiscoverX system. Importantly, compound 53
shows no activity in all three measures of D2R βꢀarrestinꢀ2 recruitment.
Despite the lack of agonist activity by 1 in the BRET assay, 1 displayed full antagonist
activity by blocking dopamineꢀstimulated βꢀarrestinꢀ2 recruitment (Figure 3E). In contrast, 1
showed partial inhibition of dopamineꢀstimulated D2R G protein activation as measured by
cAMP inhibition (Figure S2). Similarly, as observed with compound 1, compounds 24, 46, 49,
53, and 54 showed no βꢀarrestin recruitment up to 1 µM as measured by the BRET assay with
GRK2 coꢀexpressed (Figure 3F). Finally, to confirm D2R G protein activation using an
orthologous assay, we tested compounds 1 and 53 relative to aripiprazole and quinpirole using a
Gαi1ꢀGγ2 dissociation BRETꢀbased assay (Figure 3G). In this assay, compounds 1 and 53 still
exhibited potent partial agonist activity (57% and 47% of quinpirole, respectively) compared to
aripiprazole (78%). Therefore, using BRETꢀbased orthologous assay platforms for either G
protein activation or arrestin recruitment, we confirmed that compounds 1, 24, 46, 49, 53, and 54
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 63
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
are G proteinꢀbiased D2R agonists. As illustrated in Figures 2 – 3, these ligands are superior in
their D2R G proteinꢀbias profile compared to the previously reported compound 59.
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 25 of 63
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Figure 3. Compounds 1, 24, 46, 49, 53, and 54 do not activate βꢀarrestin recruitment in the D2Rꢀmediated
BRET assay while compound 59 is a partial agonist in the D2R βꢀarrestinꢀ2 recruitment Tango assay and
D2R Gi/oꢀmediated cAMP inhibition assay. (A) Activity of aripiprazole, compound 1, compound 59 and
quinpirole in the D2Rꢀmediated BRETꢀbased βꢀarrestinꢀ2 recruitment assay using HEK293 cells expressing GRK2.
Aripiprazole and compound 59 were partial agonists that promote βꢀarrestin recruitment to D2R while compound 1
was not active in the BRET assay. Quinpirole (EC50 = 7.0 nM) was used as a positive control. (B) Activity of
compound 59 and quinpirole in the D2Rꢀmediated βꢀarrestinꢀ2 translocation Tango assay using HTLA cells
transfected with a D2V2ꢀTCSꢀtTA construct. Compound 59 displayed robust βꢀarrestin recruitment in the Tango
assay (EC50 = 46 nM, Emax = 67%) compared to the full agonist quinpirole (EC50 = 2.3 nM). (C) Activity of
compound 59 and quinpirole in the D2Rꢀmediated Gi/o coupled isoproterenolꢀstimulated cAMP production assay
using HEK293T cells expressing D2R and GloSensorꢀ22F. Compound 59 was a partial agonist (EC50 = 102 nM,
Emax = 63%) compared to quinpirole (EC50 = 1.2 nM), which was used as a positive control. (D) D2R βꢀarrestin
recruitment activity by compounds 1, 53, and 59 as measured by DiscoverX in D2Rꢀexpressing CHO cells.
Compounds 1, 53, and 59 showed no activity up to 10 µM. Quinpirole (EC50 = 28 nM) was used as a positive
control. (E) Compound 1 (IC50 = 18 nM) blocked dopamine (DA, EC50 = 11 nM) stimulated βꢀarrestin recruitment
in the BRET assay. (F) Activity of compounds 24, 46, 49, 53 and 54 in the BRET assay. All compounds were not
active in the BRETꢀbased βꢀarrestinꢀ2 recruitment assay while quinpirole (EC50 = 12 nM) was used as a positive
control. (G) Activity of aripiprazole and compounds 1 and 53 in the D2R Gαi1ꢀGγ2 dissociation BRET assay.
Aripiprazole (EC50 = 0.99 nM), compound 1 (EC50 = 3.6 nM), and 53 (EC50 = 4.6 nM), all displayed partial agonist
activity (78%, 57%, and 47%, respectively) relative to quinpirole. Quinpirole (EC50 = 1.7 nM) was used as a positive
control. Except for the DiscoverX assay (n=1, performed in duplicate), data are representative of at least three
independent experiments performed in triplicate.
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
Finally, we determined binding affinity of our G proteinꢀbiased D2R agonists 1, 24, 46,
49, 53, and 54 against a panel of aminergic GPCRs and neurotransmitter transporters (Table 4).
Compound 1 exhibited tenꢀfold higher binding affinity towards D3R than D2R whereas all other
five compounds displayed similar binding affinity to D3R compared to D2R. With the exception
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 63
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 compounds 1 and 53, which had similar D4R and D2R binding affinity, all compounds
showed low or no affinity for other dopamine receptors (i.e., D1R, D4R and D5R). At serotonin
[
also known as 5ꢀhydroxytryptamine (5ꢀHT or 5HT)] receptors, all tested compounds displayed
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
moderate to high binding affinities towards 5ꢀHT , 5ꢀHT , 5ꢀHT , 5ꢀHT and 5ꢀHT . All six
1
A
2A
2B
2C
7
compounds displayed similar or higher binding affinities for 5ꢀHT and 5ꢀHT as comparing to
1
A
2B
D2R. Notably, compound 53 showed very high affinity for 5ꢀHT (Ki = 1.4 nM). High affinities
1
A
were also observed for compound 49 towards 5ꢀHT1A as well as compounds 24 and 54 towards
5ꢀHT2B (Ki < 20 nM). All compounds exhibited low or no affinities towards dopamine
transporter (DAT), serotonin transporter (SERT) and H2ꢀhistamine receptor with the exception
of compound 54 towards SERT. For H1ꢀhistamine receptor, compounds 1, 24, 53, and 54
displayed moderate binding affinities while compound 46 showed low affinity and compound 49
did not exhibit appreciable affinity.
a
Table 4. Radioligand binding affinity of compounds 1, 24, 46, 49, 53 and 54 at select GPCRs.
Binding affinity Ki (nM)
Receptor
1
24
46
170
1700
100
1800
N/A
100
N/A
820
140
71
49
53
30
54
99
D2R
D1R
D3R
D4R
D5R
ꢀHT1A
ꢀHT1B
150
N/A
13
60
45
1100
35
2000
64
2000
31
1100
69
82
280
1500
55
7100
1500
14
77
430
1800
66
N/A
86
1100
1.4
5
5
1700
320
120
38
1400
390
72
1600
94
420
100
29
2600
410
70
5ꢀHT1D
ꢀHT2A
5ꢀHT2B
ꢀHT2C
5ꢀHT3
ꢀHT5A
5
110
43
12
32
18
5
160
130
N/A
58
210
1500
1600
140
N/A
1400
92
73
330
1300
2600
330
370
810
5
ACS Paragon Plus Environment

Page 27 of 63
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
5
ꢀHT6
720
140
690
340
57
350
100
2800
450
56
960
160
2200
69
2300
31
410
95
5ꢀHT7
DAT
SERT
H1
4000
580
N/A
480
N/A
430
2500
330
52
2100
62
690
28
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
H2
680
310
2100
240
330
a
K values are the average of at least 2 duplicate experiments. SEM: < ± 20%. N/A: not active.
i
CONCLUSION
In summary, by introducing 2ꢀmethyl to the RHS benzothiazole and changing the middle
linker from propoxy to butoxy, we unexpectedly converted a βꢀarrestinꢀbiased D2R agonist to a
G proteinꢀbiased D2R agonist, compound 1. We conducted structure–functional selectivity
studies that focus on exploring two regions of the scaffold represented by this lead. These studies
revealed a number of interesting general trends and identified multiple structural motifs that
contribute to significant bias for the Gi/o signaling over βꢀarrestin recruitment. We discovered
multiple G proteinꢀbiased D2R partial agonists including compounds 24, 46, 49, 53, and 54,
which were completely inactive in two measures of βꢀarrestin recruitment, and in the case of
compound 53 inactive in all three measures of βꢀarrestin recruitment. Based on our assay results,
these newly identified ligands are superior G proteinꢀbiased D2R partial agonists compared with
the previously reported G proteinꢀbiased D2R agonist, compound 59. Among the G proteinꢀ
biased D2R agonists we have identified, compounds 24, 46 and 54 are most efficacious for the
Gi/o signaling with Emax around 40ꢀ50% and compounds 49 and 53 exhibit the most potent Gi/o
activity (EC < 10 nM). The unique bias profiles of these ligands make them a set of potentially
5
0
useful tools for elucidating the role of G proteinꢀmediated D2R signaling in health and disease.
Our studies and results will also help identify the next generation of D2R G proteinꢀbiased and βꢀ
arrestinꢀbiased ligands.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 63
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 General Procedures. Unless stated to the contrary, where applicable, the
following conditions apply: All commercial grade reagents were used without further
purification. MeCN and CH Cl were distilled from CaH under a N atmosphere before use;
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
2
2
2
THF was distilled from Na/benzophenone under N . All other dry solvents were of anhydrous
2
quality purchased from SigmaꢀAldrich. Brine (NaCl), NaHCO , and NH Cl refer to saturated
3
4
aqueous (sat aq.) solutions. Column chromatography was performed on silica gel G (200−300
mesh) with reagent grade solvents. Melting points were uncorrected. NMR spectra were recorded
1
13
on a Varian spectrometer (400 MHz for H NMR and 100 MHz for C NMR, respectively) at
1
13
ambient temperature. All H and C chemical shifts are reported in ppm (δ) relative to CDCl
3
51
(7.26 and 77.16, respectively) or CD OD (3.30 and 49.00, respectively). HPLC data for all
3
compounds were acquired using an Agilent 6110 series system with a UV detector set to 220 nm.
Samples were injected (<10 ꢁL) onto an Agilent Eclipse Plus 4.6 × 50 mm, 1.8 ꢁm, C18 column
at room temperature (rt) at a flow rate of 1.0 mL/min. A linear gradient from 10% to 100%
(vol/vol) B over 5.0 min followed by 2.0 min at 100% B with a mobile phase of (A) H O + 0.1%
2
acetic acid and (B) MeOH + 0.1% acetic acid was used. Mass spectra (MS) data were acquired in
positive ion mode using an Agilent 6110 series single quadrupole mass spectrometer with an
electrospray ionization (ESI) source. Highꢀresolution (positive ion) mass spectra (HRMS) were
acquired using a Shimadzu LCMSꢀITꢀTof timeꢀofꢀflight mass spectrometer. HPLC was used to
establish the purity of targeted compounds. All compounds that were evaluated in biological
assays had >95% purity using the HPLC methods described above.
ACS Paragon Plus Environment

Page 29 of 63
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
5
ꢀ(4ꢀ(4ꢀ(2,3ꢀDichlorophenyl)piperazinꢀ1ꢀyl)butoxy)ꢀ2ꢀmethylbenzo[d]thiazole (1). A mixture
of 2ꢀmethylbenzo[d]thiazolꢀ5ꢀol (202.7 mg, 1.23 mmol), 1,4ꢀdibromobutane (990.8 mg, 4.91
mmol) and anhydrous K CO (203.5 mg, 1.47 mmol) was dissolved in EtOH (5 mL) and the
2
3
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
solution was heated to reflux for 6 hours. The solution was diluted with water and extracted with
EtOAc. The combined organic layers were washed with saturated aq NaHCO , brine, dried over
3
anhydrous Na SO , concentrated in vacuo and purified by flash chromatography on silica gel
2
4
column to give 5ꢀ(4ꢀbromobutoxy)ꢀ2ꢀmethylbenzo[d]thiazole (25, 251.1 mg, 68% yield) as a
white solid. Compound 25 (89 mg, 0.30 mmol) was reꢀdissolved in CH CN. To this mixture was
3
added NaI (63.2 mg, 0.42 mmol) and the reaction mixture was heated to reflux for 30 min and
then cooled to rt. The commercially available compound 4 (70.8 mg, 0.31 mmol) and anhydrous
K CO (56.5 mg, 0.41 mmol) were added to the mixture. The resulting mixture was heated to
2
3
reflux and stirred for 6 h. Precipitated crystals were filtered off and the filtrate was evaporated
under reduced pressure. The residue was extracted with EtOAc. The combined EtOAc layers was
washed with brine, dried over anhydrous Na SO , concentrated in vacuo and purified by flash
2
4
chromatography on silica gel column (elution with DCM/MeOH = 50:1) to give 5ꢀ(4ꢀ(4ꢀ(2,3ꢀ
dichlorophenyl)piperazinꢀ1ꢀyl)butoxy)ꢀ2ꢀmethylbenzo[d]thiazole (1) as white solid (87.8 mg,
1
yield 65%). H NMR (400 MHz, CDCl ): δ 7.64 (d, J = 8.8 Hz, 1H), 7.44 (d, J = 2.4 Hz, 1H),
3
7.17 – 7.10 (m, 2H), 6.98 (dd, J = 8.8, 2.5 Hz, 1H), 6.95 (dd, J = 6.8, 2.8 Hz, 1H), 4.07 (t, J = 6.3
Hz, 2H), 3.08 (bs, 4H), 2.80 (s, 3H), 2.68 (bs, 4H), 2.55 – 2.48 (m, 2H), 1.93 – 1.83 (m, 2H),
13
1
1
5
.75 (ddd, J = 15.1, 8.8, 6.0 Hz, 2H); C NMR (101 MHz, CDCl ): δ 168.26, 158.32, 154.75,
3
51.36, 134.13, 127.63, 127.56, 127.40, 124.70, 121.69, 118.75, 115.15, 106.04, 68.19, 58.30,
+
3.40, 51.34, 27.32, 23.51, 20.30; HPLC: 99%, RT 4.819 min; MS (ESI) m/z 450.2 [M + H] .
+
HRMS m/z [M + H] calcd for C H Cl N OS 450.1174, found 450.1168.
2
2
26
2
3
ACS Paragon Plus Environment