Functional Characterization of a Novel Series of Biased Signaling Dopamine D3 Receptor Agonists

Article abstract of DOI:10.1021/acschemneuro.6b00221

Dopamine receptors play an integral role in controlling brain physiology. Importantly, subtype selective agonists and antagonists of dopamine receptors with biased signaling properties have been successful in treating psychiatric disorders with a low inci

Full text of DOI:10.1021/acschemneuro.6b00221

Subscriber access provided by TRENT UNIV
Article
Functional characterization of a Novel Series of
Biased Signaling Dopamine D3 Receptor Agonists
Wei Xu, Xiaozhao Wang, Aaron M. Tocker, Peng Huang, Maarten E.A.
Reith, Lee-Yuan Liu-Chen, Amos B Smith, III, and Sandhya Kortagere
ACS Chem. Neurosci., Just Accepted Manuscript • DOI: 10.1021/acschemneuro.6b00221 • Publication Date (Web): 01 Nov 2016
Downloaded from http://pubs.acs.org on November 2, 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.
ACS Chemical Neuroscience is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 47
ACS Chemical Neuroscience
1
2
3
4
5
6
Functional Characterization of a Novel Series of Biased Signaling Dopamine D3 Receptor
Agonists
7
8
9
Wei Xu1, Xiaozhao Wang3, Aaron M. Tocker1, Peng Huang5, Maarten E.A. Reith4, LeeꢀYuan Liuꢀ
Chen5, Amos B. Smith, III3, Sandhya Kortagere1,2*
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1. Department of Microbiology and Immunology, Drexel University College of Medicine, PA
19129
2. Institute for Molecular Medicine, Drexel University College of Medicine, PA, 19104
3. Department of Chemistry, University of Pennsylvania, PA, 19102
4. Department of Psychiatry, Biochemistry and Molecular Pharmacology, NYU School of
Medicine, NY, 10016
5. Department of Pharmacology, Temple University School of Medicine, PA, 19140
*Corresponding author:
Sandhya Kortagere, Ph.D.
Department of Microbiology and Immunology
Drexel University College of Medicine
2900, Queen Lane, Philadelphia, PA 19129
Tel: 215ꢀ991ꢀ8135
Email: sandhya.kortagere@drexelmed.edu
1
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 2 of 47
1
2
3
4
5
6
7
8
Keywords: biased signaling; βꢀarrestin; dopamine D3 receptors; desensitization; functional
selectivity; Gꢀproteinꢀdependent signaling; hybrid structure–based design; structure–function
relationship
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
2
ACS Paragon Plus Environment

Page 3 of 47
ACS Chemical Neuroscience
1
2
3
4
5
6
Abstract
7
8
9
Dopamine receptors play an integral role in controlling brain physiology. Importantly subꢀtype
selective agonists and antagonists of dopamine receptors with biased signaling properties have
been successful in treating psychiatric disorders with a low incidence of side effects. To this
end, we recently designed and developed SK609, a dopamine D3 receptor (D3R) selective
agonist that has atypical signaling properties. SK609 has shown efficacy in reversing akinesia
and reducing Lꢀdopaꢀinduced dyskinesia in a hemiparkinsonian rats. In the current study, we
demonstrate that SK609 has high selectivity for D3R with no binding affinity on D2R highꢀ or
lowꢀaffinity state when tested at a concentration of 10 ꢁM. In addition, SK609 and its analogues
do not induce desensitization of D3R as determined by repeated agonist treatment response in
phosphorylation of ERK1/2 functional assay. Most significantly, SK609 and its analogues
preferentially signal through the Gꢀproteinꢀdependent pathway and do not recruit βꢀarrestinꢀ2,
suggesting a functional bias towards the Gꢀproteinꢀdependent pathway. Structureꢀactivity
relationship (SAR) studies using analogues of SK609 demonstrate that the molecules bind at
the orthosteric site by maintaining the conserved salt bridge interactions with aspartate 110 on
transmembrane 3 and aryl interactions with histidine 349 on transmembrane 6, in addition to
several hydrophobic interactions with residues from transmembranes 5 and 6. The compounds
follow a strict SAR with reference to the three pharmacophore elements—substituted phenyl
ring, length of the linker connecting phenyl ring and amine group, and orientation and
hydrophobic branching groups at the amine among SK609 analogues for efficacy and functional
selectivity. These features of SK609 and the analogues suggest that biased signaling is an
inherent property of this series of molecules.
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
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 4 of 47
1
2
3
4
Introduction
5
6
7
8
9
Signaling by Gꢀprotein coupled receptors (GPCRs) has been a textbookꢀexample case study to
describe the hierarchy and utility of first, second, and third messenger systems in eliciting a
functional response. Most of this knowledge was derived based on the effects of endogenous
neurotransmitters binding to their cognate receptors and initiating signaling events that lead to
the recruitment of various coupling molecules such as Gꢀprotein subunits, second messengers,
and a cascade of kinases and phosphatases, eventually producing a cellular response 1.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Dopamine D2ꢀlike receptors consist of the subꢀtypes D2R, D3R, and D4R, which couple to the
Gi/o class of Gꢀproteins for neurotransmission. These subꢀtypes are highly homologous to each
other and have overlapping distribution and functional roles in the brain2. Hence, developing
subꢀtype selective ligands has been a longstanding challenge in the field of dopamine receptor
pharmacology. However, with newer techniques and the use of computational models in
understanding molecular determinants of function3 it is now feasible to design subꢀtype specific
molecules. In addition to subꢀtype specificity, recent trends in designing “biased functional”
ligands have added further credence to understanding the various molecular signaling
pathways4. Although GPCRs can signal via many pathways, the Gꢀproteinꢀdependent and the
Gꢀproteinꢀindependent or βꢀarrestinꢀdependent pathways have emerged as major players in
functional selectivity5. This suggests that GPCRs must exhibit a range of conformational
changes in response to biased ligands, some of which could stabilize and lead to distinct
functional effects1b, 6. Several biased signaling beta blockers and opiates have already shown
better clinical effects compared with their unbiased counterparts, suggesting that biased
signaling molecules could have better utility as clinical agents, with fewer offꢀtarget effects7.
1b, 6b
Although several crystal structures of GPCRs are available in apo or ligand bound forms
,
there is little or no structure–function correlation between the orthosteric binding pocket of the
4
ACS Paragon Plus Environment

Page 5 of 47
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
GPCRs and the functional response of the different ligand classes, namely the agonist,
antagonist, and partial or inverse agonists8. These issues are further complicated by the action
of biased ligands (agonists or antagonists) that seem to produce distinct conformational
changes resulting in functional selectivity. In fact, early references to functional selectivity aptly
described it as “functional trafficking”, as it referred not only to the unique conformational
changes in the receptor but also to the recruitment of the downstream signaling molecule9.
Hence, it is hypothesized that the signaling molecule determines the chemotypic signature of a
given agonist molecule with the functional cellular outcome10. Studies utilizing fluorescence
techniques have provided evidence for the role of functional domains with the receptor that code
for Gꢀproteinꢀdependent and ꢀindependent mechanisms in vasopressin receptors11. Accordingly,
transmembrane (TM) segment 6 and the third intracellular loop region have been shown to be
involved in Gꢀproteinꢀdependent signaling. This has been well validated by crystal structure,
nuclear magnetic resonance (NMR) studies, and molecular dynamics studies of the β2ꢀ
adrenergic receptor12. Similarly, TM7 and helixꢀ8 regions have been implicated in βꢀarrestin
recruitment and signaling of β2ꢀadrenergic receptor induced by the biased signaling agonist
carvedilol 13.
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
The D2ꢀlike receptors in their monomeric forms are traditionally known to couple to the Gi/Go
class of Gꢀproteins, initiated by a ligand binding event and resulting in activation of either the Gꢀ
proteinꢀdependent and ꢀindependent pathways or, in most cases, both pathways14. The Gꢀ
proteinꢀdependent pathway in the case of D2ꢀlike receptors leads to inhibition of adenylyl
cyclase and reduction of cAMP production, eventually leading to activation of ERK1/2 kinases.
The Gꢀproteinꢀindependent pathway is initiated by the ligand binding event, followed by
phosphorylation of the receptor by the Gꢀprotein coupled receptor kinases (GRKs), which further
leads to the recruitment of βꢀarrestinꢀ2 and activation of the glycogen synthase kinase 3β
(GSK3β)ꢀmediated pathway 15. The endogenous neurotransmitter dopamine signals through
5
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 6 of 47
1
2
3
4
both the Gꢀprotein and βꢀarrestinꢀmediated/Gꢀdependent pathways, leading to the downstream
5
6
16
effects of mitogenꢀactivated proteinꢀindependent pathways kinase (MAPK) activation
.
7
8
9
However, synthetic ligands that mimic dopamine’s signaling action of activating multiple
pathways cause undesirable side effects, which has led to the design of a new class of
dopaminergic agents that have a biased signaling profile through either the Gꢀproteinꢀ
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
dependent or βꢀarrestinꢀmediated/Gꢀproteinꢀindependent pathways16ꢀ17
.
Some known
chemotypes with D2R biased signaling profile include antipsychotic drugs such as aripiprazole
and its analogues, which recruit and signal through the βꢀarrestinꢀmediated pathway18. Similarly,
several other clinically used antipsychotic agents have been shown to signal via the βꢀarrestinꢀ
mediated pathway at the D2R, suggesting a common mechanism of action 19. Recently, Free et
20
al.
have shown a new class of biased D2R agonists that mediate Gꢀproteinꢀdependent
signaling and do not recruit βꢀarrestin, while antagonizing the dopamine stimulated βꢀarrestin
recruitment to D2R. In another interesting study on D2R by Shonberg et al, functional bias
between pꢀERK1/2 and cAMP pathways was achieved by incorporating fragments of cariprazine
to develop structureꢀactivity relationships (SARs) of a novel class of antipsychotic drugs,
suggesting that functional bias was directly linked to the chemical signature of the molecule 21.
The D3R class of receptors has been overlooked in terms of their pharmacological and
biophysical mechanisms of action owing to lack of selective agents that can distinguish the D3R
and D2Rs in vivo. However, recent evidence for their differential roles in neurological disorders
such as Parkinson’s disease (PD), Huntington’s disease, and Lꢀdopaꢀinduced dyskinesia (LID)
has spurred interest in designing selective and biased ligands of D3R22. Consequently, several
new chemotypes that possess selective activity against D3R, with a range of functional features
23
24
from agonists to antagonists , have been designed and tested in various disease models.
The crystal structure of D3R has further aided in screening and rationally designing novel small
molecule modulators of D3R25. We have recently designed novel atypical D3R agonists with
6
ACS Paragon Plus Environment

Page 7 of 47
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
biased signaling properties that differ not only from other known D2ꢀlike receptor agonists but
also from several wellꢀknown D3R agonists26. Further, the in vivo efficacy of the lead molecule
in reversing the akinesias and in blocking LIDs was recently demonstrated using a
hemiparkinsonian rodent model of PD 23a. In this study, we have further developed the SAR of
this novel series of compounds, and importantly demonstrate their functional selectivity to the Gꢀ
proteinꢀdependent and βꢀarrestinꢀindependent pathway.
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
7
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 8 of 47
1
2
3
4
Results and Discussion
5
6
7
8
Functional screening of SK609 and its analogues by ERK1/2 phosphorylation
An in silico screen was performed using the HSB method as discussed previously 26 to
derive analogues of SK609. Based on the HSB scoring scheme, 17 molecules (including
SK609) were chosen for further in vitro screening (Table 1). The molecules were tested for D2R
and D3R selectivity using the phosphorylation of ERK1/2 as an end point assay in cells stably
expressing either the D2R or D3R. DA and D2R/D3R agonist PD128907 were used as
reference compounds.
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
In CHOꢀD2R cells (Bmax 1.0 ± 0.15 pmol/mg protein), 1 ꢁM of DA and PD128907
significantly evoked ERK1/2 phosphorylation, but 1 ꢁM of SK609 and its analogues did not have
any effect (Figure 1). However, in CHOꢀD3R cells (Bmax 9.1 ± 0.4 pmol/mg protein), 1 ꢁM of DA,
PD128907, SK609, and 9 analogues significantly activated ERK1/2 phosphorylation (Figure 1).
As shown in Figure 2A, Supplementary Figure 1 and Table 2, DA, SK609, SK608, SK213, and
SK232 doseꢀdependently evoked ERK1/2 phosphorylation with EC50 values of 50.2 ± 5.9 nM,
84.6 ± 21.4, 39.2 ± 4.2 nM, and 859 ± 232 nM, respectively, at CHOꢀD3R. The Emax values for
SK609, SK608, SK213, and SK232 were 81 ± 4.5, 76.1 ± 1.4, 95 ± 5.6, and 106 ± 6.5,
respectively. Further, SK609ꢀ and analogueꢀinduced ERK1/2 activation in CHOꢀD3R cells was
completely antagonized by pretreatment with 1 ꢁM of D2ꢀlike antagonist haloperidol (Figure 2B).
Enantiomeric effect of SK609: The enantiomeric isoforms of SK609 (namely, SꢀSK609
and RꢀSK609) were tested for selectivity to D3R and functional effect on inducing ERK1/2
activation in CHOꢀD2R and CHOꢀD3R cells. D2R/D3R agonists pramipexole and PD128907,
along with the racemate SK609, were used as positive controls. The SꢀSK609 isoform was
found to be selective to D3R and had a functional effect on activation of ERK1/2 identical to that
of the racemate SK609, while RꢀSK609 had neither any effect on either receptor nor any
functional effects (Figure 3A). The doseꢀresponse curves of SꢀSK609 and RꢀSK609, along with
8
ACS Paragon Plus Environment

Page 9 of 47
ACS Chemical Neuroscience
1
2
pramipexole and racemate SK609, with increasing concentrations (10ꢀ10 to 10ꢀ4 M) were
assessed. The results confirmed the observation that SꢀSK609 had a selective and functional
D3R effect with an EC50 value of 109 ± 15.4 nM and an Emax value of 93.8 ± 14.5 (Figure 3B),
with RꢀSK609 being virtually inactive. The results suggest that the functional effect of the
racemate SK609 is due to the Sꢀenantiomer and that the Rꢀenantiomer has no effect. Indeed,
racemic SK609 displayed an EC50 value of 113.9 ± 20.3 nM (with Emax of 88.7 ± 8.6), not
appreciably different from what would be expected theoretically with S isomer being the only
active component. Based on these results, we utilized the racemate SK609 for all further
studies. In this series of experiments, pramipexole did not demonstrate selectivity to D3R
(Figure. 3A) but had a potent functional D3R effect, with an EC50 value of 2.85 ± 0.51 and an
Emax of 100.
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
Structure activity relationship of SK609 and its analogues
The structureꢀactivity relationship was determined among the 17 molecules (Tableꢀ1) using
three pharmacophore elements namely: the phenyl group; length of the linker chain connecting
the phenyl and the amine group; and the hydrophobic branching at the amine group (Table 1).
Compound SK603 has a furan ring in place of the phenyl ring, and docking studies show that
despite binding to the same pocket with the conserved salt bridge with Asp110, it had lower
binding scores compared with SK609. This could be due to the loss of aryl interactions with
His349 and other hydrophobic groups in TM5 and 6 (Figure 4E). SK603 still maintained
selectivity to D3R but had a reduction in functional effect in an ERK1/2 activation assay (Figure
1). Compounds SK058 (F), SK609 (Cl), SK032 (Br), SK182 (OH), SK052 (I), and SK041 (MeO)
were used to determine the effects of various electron withdrawing groups when substituted at
the ortho position on the phenyl ring. The order of potency was found to be in agreement with
the order of electron withdrawing properties of halogen atoms (F>Cl>Br>I) (Figure 1). However,
MeO or OH substitutions at either the para or ortho positions led to either weak efficacy (SK041)
9
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 10 of 47
1
2
3
4
5
6
7
8
9
or a loss of efficacy (SK611, SK182, and SK220) (Figure 1). In contrast, compounds SK609,
SK608, SK058, and SK610 have halogen atoms Cl or F substituted at ortho or para positions on
the phenyl ring and have the best efficacy. These positions are well known to be more efficient
than the meta position in drawing the electrons away from the pi system and increasing the
stability of these compounds. (Figure 4). This feature of halogens has been well demonstrated
among all the drugꢀlike molecules, and hence halogens have been labeled as the best
electrophilic aromatic substitutions on ring molecules 27.
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
Effect of linker length and enantiomers: The Lichtenberger compounds 28 labeled here as L1
and L2 have a para fluoro substituted phenyl ring and a primary amine group separated by a
twoꢀcarbon linker (unlike the threeꢀcarbon linker found in other analogues of SK609), thus
rendering them inactive. This short linker prevents the primary amine in compounds L1 and L2
from forming the salt bridge interaction with the aspartic acid on TM3 (Asp110) and from
simultaneously maintaining aromatic pi interactions between the phenyl ring and the aromatic
residue cluster from TM5 and TM6 (Figure 4). We also tested the effects of the R and S
enantiomers in our SAR studies because it is well known that for most dopaminergic agonists
there is a loss of efficacy with the change in orientation of the primary amine group. We found
that the Sꢀenantiomer maintained all the same interactions as those of the racemate and in fact
had better binding affinity to D3R, with additional areneꢀcation interaction with Val111 (Figure
4B). However, the Rꢀenantiomer did not form the salt bridge with Asp110 and likely does not
bind to D3R (Figure 4C). We have previously demonstrated that atypical agonists such as
SK609 induce a specific conformational change that requires simultaneous interactions of the
phenyl ring with the aromatic cluster, and a salt bridge between the primary or secondary
amines and Asp110 on TM3 of D3R. These interactions are required to induce a signature
conformational change that further locks the interactions between extracellular loops 2 and 3 in
D3R (Figure 4A)29.
10
ACS Paragon Plus Environment

Page 11 of 47
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
Effect of the hydrophobic branching group: Compounds SK220 and SK232 are derived from
parent compounds SK182 and SK609, respectively, by increasing the length or the
hydrophobicity of the branching group. SK220, similar to its parent compound SK182, is
moderately active despite gaining additional favorable interactions with Val 111 and Cys114 and
the extended hydrophobic tail. However, SK232, with an increased branching with the methanol
linker, led to an increase in activity. SK232, despite its increase in activity, could not be used as
a lead compound because we observed a loss of activity when it was stored at room
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
temperature, although it was stable when maintained at ꢀ20°C. Compound SK213 was
designed to test the effect of converting a primary amine to a secondary amine, which did not
impair the efficacy of the molecule. Docking studies showed that the molecule maintained all its
interactions with the binding site residues, similar to its parent molecule SK609.
Binding profiles of SK609 and its analogues at dopamine receptor subtypes
To further characterize the binding profiles, we chose a subset of the compounds (SK609,
SK608, SK213, SK232, SK032, SK041, SK052 and SK058) and compared them with reference
compounds DA or PD128907 and vehicle controls DMSO and/or vitamin C (VC). Binding
screening with 10 ꢁM of SK609 and its analogues was conducted using HEKꢀD1R, CHOꢀD2R,
CHOꢀD3R, HEKꢀD4R, and HEKꢀD5R cell lines (Figure 5). Results from the screening showed
that SK609 and its analogues were effective at binding to D3R but none of the test
compounds had any significant affinity for D2R (IC50>>10 ꢁM). A similar profile was observed for
D4R with the exception of SK608. However, the reference compounds DA and PD128907 had
high binding affinity for D2R, D3R and D4R in the same assay. None of the test compounds
showed binding to D1R or D5R.
Further competitive inhibition (affinity assay) by SK609 and its analogues to D3R was
conducted and the Ki values for high affinity (Ki-H) and low affinity (Ki-L) were determined. As
11
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 12 of 47
1
2
3
4
5
6
7
8
9
shown in Table 3, Ki-H values for the parent compound SK609 and the analogue SK608 were
283 ± 30 nM and 103 ± 25 nM, respectively, compared with reference compounds DA and
PD128907, with Ki-H values of 9.9 ± 4.4 and 3.7 ± 1.6 nM, respectively. The Ki-H values of other
SK609 analogues varied from 1 to 7 ꢁM. In the radioligand binding assay, SK608 exhibited
improvement in binding profile to D3R, with a Ki-H value of 103 nM compared with 283 nM for
SK609; however, neither compound had detectable binding to D2R or other DA receptors
(Figure 5 and Supplementary Figure 3), except that SK608 also had significant binding to D4R
at 10 ꢀM. Molecular docking results suggest these compounds bind at the orthosteric site with
specific interactions with residues from TM3, 5, and 626.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In order to confirm the affinity of test compounds to D2R, we reassessed the binding affinity
to CHOꢀD2R in the lowꢀ and highꢀaffinity states as described in Materials and Methods. Neither
DA and PD128907 nor any of the test compounds had affinity for D2R at 10 ꢀM (IC50>>10 ꢀM)
when CHOꢀD2R was shifted to the lowꢀaffinity state in Na+ꢀGTPγS binding buffer
(Supplementary Figure 3). However, DA and PD128907 were found to have higher affinity for
D2R than in the regular binding buffer (Figure 5), when CHOꢀD2R was shifted to the highꢀaffinity
state with pretreatment of membranes with GDP and Na+ and the binding assay conducted in
the presence of Mg2+, a finding in agreement with the studies on D2R and other GPCRs30.
However, SK609 and its analogues had no affinity to D2R even when D2R was in high affinity
state. The D2Rꢀlike antagonist raclopride had similar high affinity to D2R in both the lowꢀ and
highꢀaffinity states (Supplementary Figure 3). The results strongly support the conclusion that
SK609 and its analogues had no affinity for D2R (IC50>>10 ꢁM).
It is noteworthy that the binding assay with antagonist [3H]methylspiperone commonly used
for D2ꢀlike receptors has its limitations in determining agonist binding affinity. The antagonist
spiperone can bind both highꢀ and lowꢀaffinity sites/states of the D2Rꢀlike receptors31
. D2Rꢀlike
12
ACS Paragon Plus Environment

Page 13 of 47
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
agonists have been reported to show much lower affinity to D2R or D3R using antagonist
[3H]methylspiperone than agonist [3H]7ꢀOHꢀDPAT or other hot agonists 23b, 32. In future studies,
we will reassess and compare the binding affinities of SK609 and SK608 to D3R using
radioligand D2Rꢀlike agonists such as [3H]7ꢀOHꢀDPAT.
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
Agonistꢀinduced [35S]GTPγS binding at human dopamine D3R
Next, the compounds were tested for their ability to couple Gꢀproteins. The binding of
[35S]GTPγS to the Gα subunit of Gi/o proteins when induced by an agonist is a wellꢀknown
functional assay to characterize the potency and efficacy of an agonist. Figure 6 showed that 10
ꢁM of SK609 and its analogues increased [35S]GTPγS binding to membranes of CHOꢀD3R
cells. Among these compounds, the values of [35S]GTPγS binding increase induced by
compounds SK609, SK052, SK213, and SK232 reached 50% or more of the value of DA,
indicating that these compounds are D3R agonists and activate Gi/o proteins (Figure 6). The
potencies (EC50) and efficacies (Emax) of SK609 and its analogues and reference compounds DA
and PD128907 in stimulating CHOꢀD3R to enhance [35S]GTPγS binding are shown in Table 2.
DAꢀevoked [35S]GTPγS binding increased, with EC50 values of 11.3 ± 2.7 nM and 319.3 ± 73.1
nM at D3R and D2R, respectively, a finding in agreement with previous studies 23c, 33. Similarly,
in the same assay PD128907 had EC50 values of 1.19 ± 0.18 nM and 103.2 ± 29.6 nM at D3R
and D2R, respectively, and SK609 had an EC50 value of 1109 ± 273 nM at D3R. Compounds
SK608, SK052, SK213, and SK232 had EC50 values of 33.5 ± 8.4 nM, 1324 ± 315 nM, 1285 ±
44 nM, and 3458 ± 868 nM, respectively, at D3R. In comparison with DA (its effect defined as
100%), PD128907 had a full agonist effect on both D3R and D2R, with Emax values of 95.1 ± 2.1
and 92.8 ±1.7, respectively. In the same assay, SK609 had an EC50 value of ~1 ꢁM, with an
Emax of ~75% of DA, suggesting it may be a partial agonist at D3R; however, its analogues
SK213 and SK232 had Emax values >95% and were full agonists. In comparison, the analogue
SK608 had the strongest potency, with an EC50 value of 33.5 nM and an Emax value 81.1 ± 5.5.
13
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 14 of 47
1
2
3
4
5
6
7
8
Taken together, these results indicate that SK609 and its analogues are D3R agonists to
activate Gi/o proteins and downstream ERK1/2 signaling but have no effects on D2R/ERK1/2
signaling.
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
Time course of SK609 and its analogueꢀinduced ERK1/2 phosphorylation
In agreement with previous studies from GPCRs 18, 34, we observed that DA (1 ꢁM) elicited
both early and late phases of phosphorylation of ERK1/2 in CHOꢀD3R cells. The early phase
peaked at about 5 min and the late phase was sustained for nearly 1 h (Figure 7). However,
SK609 and its analogue SK213 (1 ꢁM) only induced an early phase (in 10 min) of ERK1/2
phosphorylation, suggesting that SK609 and its analogues may be exhibiting a biased signaling
effect to induce ERK1/2 phosphorylation, unlike DA, which is known to signal through both Gꢀ
proteinꢀ and βꢀarrestinꢀmediated pathways.
SK609 and its analogues do not induce desensitization of the D3R/ERK1/2 pathway
GPCRs undergo desensitization following prolonged or repeated activation by an agonist.
To assess whether SK609 and its analogues induce D3R desensitization compared with
reference compounds DA or PD128907, we subjected HEKꢀD3R cells to two successive
treatments with vehicle or 1 ꢁM of PD128907, SK609, and its analogues to repeatedly activate
D3R. The response to repeated activation was measured using the phosphorylation signal of
ERK1/2. As shown in Figure 8, repeated activation of D3R by PD128907 (PD/PD group)
resulted in a significant reduction of ERK1/2 phosphorylation signal compared with the control
(DMSO/PD group). However, neither SK609 nor its analogues produced a significant reduction
in phosphorylation signal, indicating that these compounds do not induce desensitization of
D3R/ERK1/2 signaling. However, other D2R/D3R agonists such as PD128907 (Figure 8) or
DA35 significantly induced D3R desensitization. Although pꢀERK1/2 is not commonly used to
14
ACS Paragon Plus Environment

Page 15 of 47
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
assess GPCR desensitization, we have demonstrated that repeated agonist treatments also
affect downstream ERK1/2 pathways for D3R, which was also demonstrated in vivo36. The
mechanisms underlying homologous desensitization of GPCRs are similar among receptors37. It
has been reported that DAꢀinduced desensitization of D3R is associated with pharmacological
sequestration in a Gβγꢀ and βꢀarrestinꢀdependent manner35b.
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
SK609 and its analogues are Gi/o/ERK1/2 biased signaling D3R agonists
To determine the biased signaling effects of SK609 and its analogues, we tested their
effects on βꢀarrestinꢀ2 recruitment after D3R activation. We used the PathHunter eXpress
(DiscoverX, Fremont, CA) U2OS D3R βꢀarrestin recruitment assay, which uses a βꢀ
galactosidase fragment complementationꢀbased technology to monitor the interaction of βꢀ
arrestin with GPCRs. As shown in Figure 9, while PD128907 potently evoked D3Rꢀmediated βꢀ
arrestin recruitment, with an EC50 of 5.5 ± 0.18 nM, SK609 and its analogue SK213 had very
weak effects, with EC50 values computed as 14.4 ± 2.7 ꢁM and 6.8 ± 0.5 ꢁM, respectively, and
Emax values of 57.3 ± 2.5 and 66.1 ± 6.5 (% of PD128907), respectively; the latter values are
perhaps underestimates, obtained from incomplete curves that stop at the high practical limit of
100 uM (not reaching the plateau) (Figure 9).
A comparison of the effect of SK609 and its analogues across all the functional
pathways is shown in Table 2. The ratios of the EC50 values of SK609 and SK213 at βꢀarrestin
recruitment and Gꢀproteinꢀdependent ERK1/2 phosphorylation are 286 or 173, respectively; the
ratios of Emax values of SK609 and SK213 for βꢀarrestin recruitment and Gꢀproteinꢀinduced
ERK1/2 phosphorylation are 0.70 or 0.69, respectively. The ratios of EC50 values of SK609 and
SK213 for Gi/o coupling and the ERK1/2 pathway are 21 or 33, respectively; the ratios of Emax
values of SK609 and SK213 are 0.91 and 1.03, respectively. Taken together, these results
indicate that SK609 and SK213 have higher potency for Gꢀproteinꢀinduced ERK1/2
phosphorylation, with a low tendency to recruit βꢀarrestin, and are Gi/o/ERK1/2 biased signaling
15
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 16 of 47
1
2
3
4
agonists of D3R.
5
6
7
8
9
Screening for offꢀtarget effects of SK609 using βꢀarrestin recruitment assay
A family of GPCRs (128 receptors) was screened using DiscoverX using their platform βꢀ
arrestin recruitment technology in both agonist and antagonist mode. At a high concentration
(100 ꢁM), SK609 had significant affinities only to D2R, D3R, CB1, 5HT2A, and β2ꢀadrenergic
receptors, and showed agonist activities at D2R and D3R and antagonist activity at β2ꢀ
adrenergic, cannabinoid, and 5HT2A receptors (Figure 10A). Further testing in a 10ꢀpoint doseꢀ
response study confirmed that neither SK609 nor its analogue SK213 had antagonizing effects
on the isoproterenolꢀinduced β2ꢀadrenergic receptor activation, with an IC50>10 ꢀM (Figure
10B).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In conclusion, the results obtained indicate that SK609 and its analogues are novel
biased signaling agonists of D3R. These compounds have no affinity to D2R and do not
significantly bind to other GPCRs. They possess unique signaling features in that they do not
recruit βꢀarrestinꢀ2 for downstream signaling and they show specificity towards Gꢀproteinꢀ
dependent signaling to activate and phosphorylate ERK1/2. In addition to being biased signaling
agonists, SK609 and its analogues also possess atypical agonist features that are not
commonly known among the D3R agonists. SK609 and its analogues do not induce receptor
desensitization as demonstrated with the ERK1/2 phosphorylation assay with repeated agonist
treatment. Our results demonstrate that SK609 and its analogues do not significantly recruit βꢀ
arrestin and hence do not induce desensitization. These observations are in agreement with our
previous reports that SK609 does not induce tolerance at the GIRK channels 26. We have
recently demonstrated that SK609 has very potent effects in reversing akinesia in a
hemiparkinsonian rodent model 23a. The unique signaling features of these molecules (including
no tolerance and desensitization) are also reflected in the in vivo effects of SK609, in which
16
ACS Paragon Plus Environment

Page 17 of 47
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
chronic treatment with the compound does not lead to loss of efficacy as is seen with most other
D3Rꢀselective agents, such as pramipexole38. In addition to its ability to reverse akinesia,
SK609 can be combined with Lꢀdopa synergistically to improve motor deficits without
aggravating the dyskinesias. In fact, SK609 can block the dyskinetic effects produced by Lꢀ
dopa, an action that can be attributed to the partial agonistic property of SK609 at postsynaptic
D3R present mostly as D1RꢀD3R heteromers 39. Our future efforts will be geared towards
understanding the functional effects of biased signaling agonists such as SK609 on D1R/D3R
cross talk at the membrane and their in vivo effects in naïve and hemiparkinsonian rodent
models.
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
Materials and Methods
Materials
[3H]methyl spiperone (85.5 Ci/mmol), [3H]SCH23390 (73.1 Ci/mmol), and [35S]GTPγS
(1250 Ci/mmol) were purchased from PerkinElmer (Boston, MA). Tetracycline, hygromycin,
blasticidin, dopamine (DA), (+) butaclamol, fluphenazine, phenylmethylsulfonyl fluoride (PMSF),
GDP, GTPγS, and ascorbic acid were obtained from SigmaꢀAldrich (St. Louis, MO). G418 was
obtained from Gemini BioꢀProducts (West Sacramento, CA); cell culture reagents were obtained
from Invitrogen (Carlsbad, CA); and protease inhibitor cocktail, phosphatase inhibitor cocktail,
and lysis buffer were obtained from Thermo Scientific (Rockford, IL). All test molecules (Table 1)
were either synthesized inꢀhouse as described below or purchased from vendors: Asinex Corp.
(WinstonꢀSalem, NC): SK603, SK609, SK610, SK611; Enamine, Ltd. (Kiev, Ukraine):
Compounds described in Lichtenberger patent 28 ꢀ L1 and L2; Asiba Pharmatech, Inc. (Milltown,
NJ): R and S enantiomers of SK609. DA and D2R/D3R agonists PD128907 and pramipexole
were purchased from Tocris Bioscience (BioꢀTechne, Inc., CA).
17
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 18 of 47
1
2
3
4
5
6
7
8
9
Synthesis of SK609 and its Analogues: The preparation of a series of SK609 analogues was
achieved through exploration of different synthetic routes. For example, synthesis of SK032•HCl
began with an alkylation/deꢀacylation cascade, using starting material benzylbromide 1b
(Scheme 1). Reductive amination was carried out using NaBH3CN as reducing agent, to afford
the desired product SK032, which was then converted to its HCl salt. Compounds SK058•HCl
and SK052•HCl, both as HCl salts, were prepared following the same synthetic sequence. This
route can also be applied to deliver compound SK041•HCl, which was then treated with HBr in
H2O to generate the phenol analogue SK182 (Scheme 2).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
To prepare compound SK213•HCl, the alkylation/deꢀacylation sequence was performed on
benzylchloride starting material 6 to afford ketone 7 (Scheme 3). Reductive amination reaction
employing MeNH2•HCl then produced a secondary amine SK213 in good yield (75%), which
again was converted to the HCl salt using HCl in Et2O solution.
To access compound SK220, benzylchloride 8 was first alkylated with βꢀketo ester 9, with NaH
as base, and in the presence of tetrabutylammonium iodide (TBAI) (Scheme 4). The resulting
ester 10 was then heated in EtOH and H2O under basic conditions (KOH) to hydrolyze the ethyl
ester, followed by decarboxylation to deliver ketone intermediate 11 in 51% yield over two steps.
The carbonyl moiety in 11 was then converted to the corresponding primary amine using the
method as described above (NH4OAc and NaBH3CN). Finally, demethylation of compound 12
with HBr in H2O was achieved, followed by highꢀperformance liquid chromatography (HPLC)
purification, to afford phenol analogue SK220.
The synthesis of primary hydroxyl analogue SK232•HCl utilized the bromo starting material 13,
which was first alkylated with diester 14 (Scheme 5). The diester 15 that was thus generated
was then treated with LiBr/H2O in DMF to produce intermediate 16. The ester functionality in 16
18
ACS Paragon Plus Environment

Page 19 of 47
ACS Chemical Neuroscience
1
2
3
4
was reduced to alcohol with LiAlH4, and the Boc protecting group was subsequently removed to
5
6
deliver the desired product SK232•HCl.
7
8
9
Finally, compound SK609 was commercially purchased as a racemate and converted to its HCl
salt using the previously described procedure, without complications (Scheme 6). Detailed
experimental procedures and characterization data (1H NMR, 13C NMR, and highꢀresolution
mass spectrometry [HRMS]) of all the aboveꢀmentioned compounds has been described in the
Supporting Information section.
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
Enantiomeric forms: The enantiomeric forms of SK609 were purchased from Asiba
Pharmatech and were synthesized in an asymmetric fashion with an enantiomeric enrichment of
~98%. The enantiomer enrichment was performed using chiral HPLC as analytical method and
diastereomeric salt formation or isomer separation as an enrichment method. The enantiomeric
forms and purity were confirmed using liquid chromatography–mass spectrometry (LCꢀMS)
(Supplementary Figure 2).
In silico screening and molecular docking
We have previously reported on differences in receptor conformations induced by tolerance
producing and nonꢀtolerance producing agonists at the D3R 26, 29. In addition, using iterative
rounds of siteꢀdirected mutagenesis and modeling studies, we have demonstrated the role of
the extracellular loopꢀ2 (EL2) and TM3ꢀIL2 regions in contributing to a tolerance phenotype in
quinpirole and PD128907 29, 40. Using this information and the pharmacophore of cisꢀ8ꢀOHꢀPBZI
(a nonꢀtolerance producing D3R agonist) as template to the hybrid structure based (HSB)
method of in silico screening, we identified SK609 as the bestꢀranking molecule that displayed
25a
high selectivity for D3R 26. The crystal structure of D3R
was prepared by adding hydrogen
19
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 20 of 47
1
2
3
4
5
6
7
8
atoms and setup in a POPC (1ꢀpalmitoylꢀ2ꢀoleoylꢀsnꢀglyceroꢀ3ꢀphosphocholine) membrane
environment. Further the membrane bound D3R was refined with 10 ns of molecular dynamics
simulation using the NAMD (Nanoscale Molecular Dynamics) program and Amber99 (Assisted
9
41
Model Building with Energy Refinement) charges and force field . We utilized the three major
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
pharmacophore features of SK609 to derive analogues using the HSB method. The features
included an aromatic ring with electrophilic substituents, a secondary or tertiary amine group,
and a threeꢀcarbon linker between the aromatic ring and the amine group. The best 50 hits that
conformed with respect to at least two features (aromatic ring and amine group) and ranked by
Tanimoto similarity score (similarity to SK609) were chosen for docking studies. The molecules
26
were docked to the previously characterized binding site consisting of residues from TM3, 5,
and 6 and EL2 using GOLD program (ver 4.1) 42. The complexes were scored using a hybrid
scoring scheme 43, consisting of weighting the GOLD score with a custom score that was
derived by scoring ring interactions with residues from TM5 and 6 and amine interactions with
Asp110 and penalizing reversed poses 43. The bestꢀranking molecules and a few worstꢀranking
molecules (docked with the correct pose but without interactions with Asp110, such as Rꢀ
SK609, L1, and L2 molecules) were chosen for further in vitro validation.
Cell Culture
HEKꢀ293 cells stably expressing human DA D1 receptor (HEKꢀD1R), D2 receptor (HEKꢀ
D2R), D3 receptor (HEKꢀD3R), D4 receptor (HEKꢀD4R), or D5 receptor (HEKꢀD5R) and CHO
cells stably expressing human D2 receptor (CHOꢀD2R) or human D3 receptor (CHOꢀD3R) were
grown in 100ꢀmm culture dishes in Dulbecco’s modified Eagle medium (DMEM) supplemented
with 10% fetal calf serum, 100 units/mL penicillin, and 100 ꢂg/mL streptomycin in a humidified
atmosphere consisting of 5% CO2 and 95% air at 37°C. HEKꢀD1R, HEKꢀD2R, or HEKꢀD3R
cells, which have the CMV promoter controlled by a Tetꢀon mechanism, were grown in the
20
ACS Paragon Plus Environment

Page 21 of 47
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
presence of 50 ꢂg/mL hygromycin and 5 ꢂg/mL blasticidin to maintain Tꢀrex and receptor
selection, respectively. The HEK cells were induced to express D1R, D2R, or D3R with 1 ꢂg/mL
tetracycline for 12 to 24 h before binding or functional assays. CHOꢀD2R, CHOꢀD3R, HEKꢀD4R,
or HEKꢀD5R cells were grown in the presence of 0.1 mg/mL G418 or 5 ꢂg/mL blasticidin to
maintain high receptor expression levels.
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
Receptor binding
Dopamine receptor binding was performed according to our published procedure 33,33
For D1R and D5R, [3H]SCH23390 was used as the radiolabeled ligand and
.
fluphenazine (10 ꢂM) was used to define nonspecific binding, whereas [3H]methylspiperone and
(+)ꢀbutaclamol (4 ꢂM) were used for radiolabeling and nonspecific binding, respectively for D2R,
D3R, and D4R. Membranes were prepared from transfected CHOꢀD2R and CHOꢀD3R cells or
HEKꢀD1R, HEKꢀD5R, HEKꢀD2R, HEKꢀD3R, and HEKꢀD4R cells as described previously33.
Saturation binding of [3H]methylspiperone to D2R and D3R was performed with at least 8
concentrations of [3H]methylspiperone (ranging from 20 pM to 10 nM). Binding was carried out
in 50 mM TrisꢀHCl buffer containing 120 mM NaCl, 5 mM KCl, 2 mM CaCl2, and 1 mM MgCl2
(pH 7.4) at room temperature for 1 h in duplicate in a volume of 250 ꢂL with 10 to 100 ꢂg
membrane protein depending on receptor expression level. Incubations were terminated by
filtration through Whatman GF/B filters (GE Healthcare Lifesciences) under vacuum, and
radioactivity on filters was measured. Competitive inhibition of [3H]methylspiperone (1 nM)
binding to D2R or D3R by test compounds and positive controls DA and PD128907 was
performed with varying concentrations (10ꢀ11 to 10ꢀ5 M) of test compounds. Binding data was
analyzed with the Prism program (GraphPad, San Diego, CA), Kd and Bmax values were
determined, and further Ki values were determined using the oneꢀsite or twoꢀsite analysis. A
similar analysis could not be performed for D2R because the compounds had no affinity to D2R
21
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 22 of 47
1
2
3
4
5
6
7
8
at 10 ꢂM. To further distinguish the effect of compounds on D2Rs’ highꢀ and lowꢀaffinity states
44, two additional binding buffers were used to determine the affinity of test compounds in
competing for the binding of [3H]methylspiperone to CHOꢀD2R. The Na+ꢀGTPγS binding buffer
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
(50 mM TrisꢀHCl, pH 7.4, with 154 mM NaCl, 0.025% ascorbic acid, 10 ꢀM GTPγS, and 0.001%
bovine serum albumin), which was used to convert the receptors to a lowꢀaffinity state for
agonists30a and the Mg2+ binding buffer (50 mM TrisꢀHCl, pH 7.4, 6 mM MgCl2, 0.025% ascorbic
acid, and 0.001% bovine serum albumin) was used for highꢀaffinity state agonist binding. In
addition to using the Mg2+ buffer, the membranes were preꢀincubated with 100 ꢀM GDP and 100
mM NaCl for 30 min at room temperature and washed and centrifuged at 40,000 × g for 15 min
to remove excess Na+ and GDP from membranes, which was used to promote the receptors to
a highꢀaffinity state for agonists in the presence of Mg2+30b, 45. [3H]methylspiperone (about 0.25
nM) binding to D2R was conducted with vehicle or 10 ꢀM of DA and PD128907 and D2Rꢀlike
antagonist raclopride (RAC) or test compounds for 1 h at 30°C in the two binding buffers,
respectively. Haloperidol 10 ꢀM was used to define nonspecific binding.
[35S]GTPγS binding assay
Determination of [35S]GTPγS binding to G proteins was carried out using a modification
of our published procedure 33, 46. Briefly, membranes of CHOꢀD2R or CHOꢀD3R cells were
prepared using published procedures 47.47. Membranes (20 ꢂg protein) were incubated in
reaction buffer containing [35S]GTPγS (80–100 pM) and 3 or 6 ꢂM GDP, 10 or 3 mM MgCl2, and
150 or 100 mM NaCl with various concentrations of each test compound (10ꢀ10 to 10ꢀ4 M) in a
total volume of 0.5 mL for 60 min at 30°C. Nonspecific binding was determined in the presence
of 10 ꢂM GTPγS. Bound and free [35S]GTPγS were separated by filtration. Data was analyzed
and EC50 and Emax values were determined with the Prism program.
22
ACS Paragon Plus Environment

Page 23 of 47
ACS Chemical Neuroscience
1
2
3
4
Immunoblot analysis of phosphorylated ERK1/2 MAPK
5
6
7
8
CHOꢀD3R and CHOꢀD2R cells (0.3×106) or HEKꢀD2R and HEKꢀD3R cells (0.6×106) were
seeded in 6ꢀwell plates and cultured in 3 mL DMEM complete growth medium for 48 h. Cells
were washed with 1 mL phosphateꢀbuffered saline (PBS) and incubated with 3 mL prewarmed
DMEM without serum for 2 h. Reference compound DA (1 ꢁM with 10 ꢁg/mL ascorbic acid) or
D2Rꢀlike agonist PD128907 (1 ꢁM in 0.01% dimethyl sulfoxide [DMSO]) and test compounds (1
ꢁM in 0.01% DMSO) were added to the cells and incubated for the indicated times at 37°C. The
reaction was stopped by aspirating the medium and washing the cells with 1 mL PBS followed
by the addition of 100 ꢁL/well lysis buffer with 1× Protease Inhibitor Cocktail, 1× Phosphatase
Inhibitor Cocktail, and 1 mM PMSF to solubilize cells. Cell lysates were assayed for protein
content using the DC Protein Assay Kit II (BioꢀRad). Loading buffer was added to protein
samples and boiled for 5 min at 100°C in a water bath, and spun at 2500 g. 5 ꢂg total
protein/lane was loaded to Novex 4ꢀ20% TrisꢀGlycine Mini Gel for separation (Invitrogen) for 2 h
and transferred to a PVDF membrane (Thermo Scientific). Immunoblotting was performed with
rabbit antiphosphoꢀp44/42 MAPK polyclonal antibody (1:5,000, Cell Signaling Technology) and
peroxidaseꢀconjugated goat antirabbit IgG (H+L) (1:5,000, Jackson Immuno Research
Laboratories, Inc.,). Total p44/42 MAPK levels in the same blots were also determined with
rabbit antip44/42 MAPK monoclonal or polyclonal antibody (1:5000, Cell Signaling Technology)
after stripping the blots of the same PVDF membrane. Chemiluminescence detection was
performed using the SuperSignal West Dura Extended Duration Substrate detection kit (Thermo
Scientific). Phosphorylated or total ERK1/2 MAPK immunoblots were quantified by densitometry
with ImageQuant LAS4000 (GE Healthcare BioꢀSciences, Pittsburgh, PA) and normalized to
vehicle treatment. Data was analyzed and EC50 and Emax values were determined with the Prism
program.
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
For desensitization experiments, 0.6×106 HEKꢀD3R cells were seeded in 6ꢀwell plates and
23
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 24 of 47
1
2
3
4
5
6
7
8
9
cultured in 3 mL DMEM complete growth medium for 48 h. Cells were washed with 1 mL PBS
and incubated with 3 mL prewarmed DMEM without serum for 2 h. Vehicle, 1 ꢁM of PD128907,
or test compounds were added to the cells and incubated for 30 min (first time pretreatment) at
37°C in CO2 incubator. The reaction was stopped by aspirating the medium and the cells were
washed 4 times (5 min/per wash in CO2 incubator) with 3 mL DMEM without serum and
subsequently treated a second time with vehicle or 1 ꢁM of PD128907, test compounds for 2
min at 37°C. The reaction was stopped by aspirating the medium and the cells were washed
once with 1 mL PBS followed by the addition of 100 ꢁL/well lysis buffer with 1× Protease
Inhibitor Cocktail and 1× Phosphatase Inhibitor Cocktail and 1 mM PMSF to solubilize cells for
immunoblot assays.
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 Recruitment DiscoveRx Assay
The measurement of βꢀarrestin recruitment stimulated by D3R or β2ꢀadrenergic receptor
activation was conducted using D3R U2OS or ADRB2ꢀCHOꢀK1 PathHunter express cells
purchased from DiscoveRx (Fremont, CA) according to the manufacturer’s protocol. In brief,
cells were thawed, resuspended in the supplied medium, plated at 10,000 cells/well in a 96ꢀwell
plate, and incubated for 48 h at 37°C. After stimulation with vehicle or various concentrations
(10ꢀ10 to 10ꢀ4 M) of test compounds in triplicate for 90 min at 37°C, the detection reagent solution
was added and incubated for 60 min at room temperature. For an antagonizing effect of test
compounds on isoproterenolꢀinduced βꢀarrestin recruitment, ADRB2ꢀCHOꢀK1 cells were
pretreated with vehicle or various concentrations (10ꢀ9 to 10ꢀ4 M) of test compounds for 30 min at
37°C. Further, vehicle or 300 nM of isoproterenol (EC80) was added and incubated for 90 min at
37°C, and the detection reagent solution was then added. Chemiluminescence was recorded
using Fluoroskan Ascent FL (Thermo Fisher Scientific). Data was analyzed and EC50 or IC50 and
Emax values were determined using the Prism program. The βꢀarrestin recruitment assay was
24
ACS Paragon Plus Environment

Page 25 of 47
ACS Chemical Neuroscience
1
2
3
4
5
6
also used to assess the effect of SK609 in a GPCR panel assay consisting of 124 receptors in
both agonist and antagonist mode at a high concentration of 100 ꢁM.
7
8
9
Determination of the Protein Content: Protein content in the membrane was determined
using the DC Protein Assay Kit II with bovine serum albumin as the standard (BioꢀRad,
Hercules, CA), following the manufacturer’s protocol.
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
Statistical Analysis
All data is presented as the mean ± s.e.m. unless specified otherwise. For comparison of
multiple groups, data was analyzed using oneꢀway ANOVA followed by a NewmanꢀKeuls
multipleꢀcomparison test. For comparison of two groups, Student's t test was performed. P<0.05
was defined as a statistically significant difference.
Supplementary information: The accompanying supplementary information file consists of
additional synthesis and HNMR profiles of all the compounds synthesized in our laboratory and
described under schemes 1ꢀ6. The LCꢀMS profiles of R, S and the racemate forms of SK609
are also listed. Supplementary figure 2 is a western blot image associated with Figure 2A and
supplementary figure 3 shows the comparison of binding affinity of SK609 and its analogues to
CHOꢀD2R in the low affinity state and the high affinity state.
Acknowledgments: SK acknowledges funding support from Drexel University Innovation fund,
DrexelꢀCoulter translational research grants and faculty development funds from Department of
Microbiology, Drexel University. AT was funded by a Drexel University undergraduate summer
research program fellowship. LYLC acknowledges funding from NIH P30 DA013429
.
25
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 26 of 47
1
2
3
4
5
6
Conflict of Interest Statement: All molecules listed in this manuscript are currently listed under a
published patent: US20140228355 and licensed to Polycore Therapeutics LLC.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 1: Structure–activity relationship profiles of SK609 and its analogues along with their
docking scores at D3R
R1
R1
R2
R1
Name
R1
R2
26
Hybrid docking score
ACS Paragon Plus Environment

Page 27 of 47
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
SK609*
SK609-R
SK609-S
SK611
o-Cl
76.8
40.3
81
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
o-Cl
o-Cl
p-OCH3
57.2
72.6
83
SK610
p-F
SK608
p-Cl
SK213
o-Cl
73.7
27
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 28 of 47
1
2
3
4
5
6
7
8
SK220
SK052
SK058
SK232
SK182
SK032
SK041
o-OH
56.2
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
o-I
65
o-F
75.9
o-Cl
78
o-OH
o-Br
o-OCH3
54.8
64.4
62
28
ACS Paragon Plus Environment

Page 29 of 47
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
L1
L2
o-F
o-F
28
48
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
*SK603 (4-(fluran-2-yl)butan-2-amine) was structurally similar to SK609 but had a furan ring instead of
an ortho chlorophenyl ring. The hybrid docking score was 64.8.
Table 2: Functional profiles of SK609 and its analogues for Gꢀproteinꢀdependent and βꢀ
arrestinꢀdependent signaling mechanisms at D3R
[35S]GTPγS binding
βꢀArrestin recruitment
EC50
Compound
DA
ERK1/2 phosphorylation
EC50
11.3 ± 2.7 nM
Emax
100
EC50
Emax
Emax
ND
5.2 ± 1.0 nM
100
ND
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
PD128907 1.2 ± 0.18 nM 96.1 ± 2.1
7.1 ± 0.5 nM
50.2 ± 5.9 nM
39.2 ± 4.2 nM
859 ± 232 nM
84.6 ± 21.4 nM
100
5.5 ± 0.18 nM
14.4 ± 2.7 ꢀM
6.8 ± 0.5 ꢀM
ND
100
57.3 ± 2.5
66.1 ± 6.5
ND
SK609
SK213
SK232
SK608
74.3 ± 2.3
98.4 ± 7.9
95.6 ± 8.1
81.1 ± 5.5
81 ± 4.5
95 ± 5.5
106 ± 6.5
76 ± 1.4
1.1 ± 0.2 ꢀM
1.3 ± 0.3 ꢀM
3.5 ± 0.9 ꢀM
33.5 ± 8.4 nM
ND
ND
For ERK1/2 phosphorylation, CHOꢀD3R cells were treated with DA or test compounds
for 5 min at 37
°C.
ND, not determined.
Table 3. Ki values (nM) of SK609 and its analogue or reference compound binding to the
CHOꢀD3R: Competitive inhibition by SK609 and its analogues or reference compounds of
29
ACS Paragon Plus Environment