Hydroxy-Substituted Heteroarylpiperazines: Novel Scaffolds for β-Arrestin-Biased D2R Agonists

Article abstract of DOI:10.1021/acs.jmedchem.7b00363

By means of a formal structural hybridization of the antipsychotic drug aripiprazole and the heterocyclic catecholamine surrogates present in the β₂-adrenoceptor agonists procaterol and BI-167107 (4), we designed and synthesized a collection of novel hydroxy-substituted heteroarylpiperazines and heteroarylhomopiperazines with high dopamine D₂ receptor (D₂R) affinity. In contrast to the weak agonistic behavior of aripiprazole, these ligands are capable of effectively mimicking those interactions of dopamine and the D₂R that are crucial for an active state, leading to the recruitment of β-arrestin-2. Interestingly, some ligands show considerably lower intrinsic activity in guanine nucleotide exchange experiments at D₂R and consequently represent biased agonists favoring β-arrestin-2 recruitment over canonical G protein activation. The ligands' agonistic properties are substantially driven by the presence of an endocyclic H-bond donor.

Full text of DOI:10.1021/acs.jmedchem.7b00363

Subscriber access provided by Eastern Michigan University | Bruce T. Halle Library
Article
Hydroxy-Substituted Heteroarylpiperazines: Novel
2
Scaffolds for #-Arrestin-Biased DR Agonists
Barbara Männel, Daniela Dengler, Jeremy Shonberg, Harald Hübner, Dorothee Möller, and Peter Gmeiner
J. Med. Chem., Just Accepted Manuscript • Publication Date (Web): 10 May 2017
Downloaded from http://pubs.acs.org on May 10, 2017
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 68
Journal of Medicinal Chemistry
1
2
3
HydroxyꢀSubstituted Heteroarylpiperazines:
4
5
6
7
Novel Scaffolds for βꢀArrestinꢀBiased D2R Agonists
8
Barbara Männel, Daniela Dengler, Jeremy Shonberg, Harald Hübner,
Dorothee Möller, Peter Gmeiner*
9
10
11
12
13
14
Department of Chemistry and Pharmacy, Medicinal Chemistry, FriedrichꢀAlexander University
ErlangenꢀNuernberg, Schuhstraße 19, 91052 Erlangen, Germany
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
ABSTRACT
By means of a formal structural hybridization of the antipsychotic drug aripiprazole and the
heterocyclic catecholamine surrogates present in the β2ꢀadrenoceptor agonists procaterol and BIꢀ
167107 (4), we designed and synthesized
a
collection of novel hydroxyꢀsubstituted
heteroarylpiperazines and heteroarylhomopiperazines with high dopamine D2 receptor (D2R) affinity.
In contrast to the weak agonistic behavior of aripiprazole, these ligands are capable of effectively
mimicking those interactions of dopamine and the D2R that are crucial for an active state, leading to the
recruitment of βꢀarrestinꢀ2. Interestingly, some ligands show considerably lower intrinsic activity in
guanine nucleotide exchange experiments at D2R and consequently represent biased agonists favoring
βꢀarrestinꢀ2 recruitment over canonical G protein activation. The ligands’ agonistic properties are
substantially driven by the presence of an endocyclic Hꢀbond donor.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 68
1
2
3
INTRODUCTION
4
5
6
7
8
The five different dopamine receptor subtypes (D1RꢀD5R) are widely distributed throughout the
body, especially within the central nervous system (CNS). These G protein coupled receptors (GPCRs)
mediate the physiological functions of dopamine, a neurotransmitter and hormone with a
catecholamine structure.1 Nonꢀsurprisingly, malfunctions of the dopaminergic neurotransmission are
involved in a number of pathophysiological conditions including schizophrenia, addiction and
Parkinson’s disease (PD). PD is characterized by severe locomotor deficits, which are mainly caused
by the loss of dopaminergic neurons in the substantia nigra pars compacta. Besides the cardinal
symptoms akinesia, rigidity and tremor at rest, patients also exhibit cognitive and other nonꢀmotor
symptoms in both, advanced and early stages of the disease.2 Despite considerable efforts to develop
novel therapeutic agents, LꢀDOPA (Lꢀ3,4ꢀdihydroxyphenylalanine) applied together with a peripheral
decarboxylase inhibitor remains the ‘gold standard’ in symptomatic antiparkinsonian therapy.
However, the prolonged use of this dopamine precursor results in a loss of efficacy over time, resulting
first in “offꢀphases” and concomitantly occurring LꢀDOPAꢀinduced dyskinesias (LID) during the “onꢀ
phases” at later stages of the disease.3 Attempts to improve fluctuations or to delay the onset of
fluctuations and LIDs with D2R agonists as monotherapy or in combination with a lower LꢀDOPA dose
did not prove successful, as the onset of LID is not dependent on the duration of therapy, but rather the
duration of the disease.4, 5 The exact mechanisms underlying LIDs remain a matter of debate and
probably depend on the overall context of denervation and compensatory processes in the CNS.
However, animal models of PD suggest LIDs may be associated with uncontrolled neuronal excitability
and dopamine receptor supersensitivity.6, 7
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
Activation of D2Rs, as for a plethora of other GPCRs, leads to conformational changes
transmitted from the ligandꢀbinding pocket towards intracellular receptor domains. Distinct receptor
conformations are thought to guide the intracellular signaling events by interaction of the receptor with
ACS Paragon Plus Environment

Page 3 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
different signal transducers, namely G proteins and βꢀarrestinꢀ2.8 Using evolutionary trace for the
generation of D2R mutants specifically interacting with the G protein or βꢀarrestin signaling pathway, it
was demonstrated that G proteinꢀ and βꢀarrestinꢀmediated signaling at the D2R is dissociable and has
distinct physiological outcomes. Although the generated mutant receptors differed in only two residues
compared to wildꢀtype, only the βꢀarrestinꢀspecific D2R mutant was able to promote significant
amphetamineꢀinduced forward locomotion in mice.9 Preferential activation of one signaling pathway
over the other does not only occur in response to genetically modified receptors but also in a ligandꢀ
dependent fashion. This ligandꢀdirected signaling is known as ‘functional selectivity’, ‘ligand bias’ or
‘biased agonism’ and has gained considerable interest over the past decades.10, 11
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
Very recently, it was demonstrated in various animal models of PD, that dopamine receptorꢀ
mediated activation of βꢀarrestinꢀ2 is required for the therapeutic effects of LꢀDOPA. In contrast,
genetic deletion of the signal transducer molecule markedly enhanced dyskinesiaꢀlike effects of both,
acute and chronic LꢀDOPA treatment and significantly reduced forward locomotion.12 Similar results
were obtained when GRK6, a kinase upstream in the βꢀarrestinꢀ2 signaling cascade, was deleted.13
Noteworthy, overexpression of βꢀarrestinꢀ2 in PD animal models enhanced the therapeutic effects of Lꢀ
DOPA, while alleviating dyskinesia.12 Specifically targeting the D2Rꢀmediated activation of βꢀarrestinꢀ
2 might therefore be a valuable approach for the design of novel antiparkinsonian drugs. Here we
report on structural modifications of aripiprazole, a weak D2R partial agonist, leading to novel βꢀ
arrestinꢀbiased D2R agonists. Previously, a number of βꢀarrestinꢀ2 biased partial agonists for D2Rs
including prototypic ligands 1 (UNC0006)14, 2 (UNC9975)14 and 3 (UNC9994)14 have been described
as putative novel antipsychotics (Chart 1). However, these ligands displayed relatively weak intrinsic
activities, and none of them exhibited dopamineꢀlike properties for the recruitment of βꢀarrestinꢀ2.
Unlike these partial agonists, our results show hydroxyꢀsubstituted heteroarylpiperazines capable to
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 68
1
2
3
4
5
fully activate the D2Rꢀpromoted recruitment of βꢀarrestinꢀ2, while displaying considerably lower
tendency to stimulate G protein mediated signaling.
6
7
8
Chart 1. Previously described βꢀarrestinꢀbiased agonists 1ꢀ3 with partial agonist characteristics at D2R.
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
RESULTS AND DISCUSSION
Ligand design. When aiming to develop functionally selective agonists, a ligandꢀbased strategy
suggests starting from a balanced agonist and trying to replace or remove molecule parts associated
with receptor activation. If such a modification disrupts one or the other signaling pathway more
effectively, ligand bias will be observed. A complementary design embarks on an antagonist scaffold,
and depends on the inclusion of agonistꢀtype structural moieties. Thus, biased agonists may be accessed
by designing hybridꢀstructures containing agonist and antagonist moieties. Starting from the typical
1,4ꢀdisubstituted aromatic piperazine (1,4ꢀDAP) aripiprazole, a typical weak D2R partial agonist, we
envisioned incorporation of typical catecholꢀlike agonist substructures in order to generate putatively
biased D2R agonists. The head groups of typical 1,4ꢀDAPs are not appropriately substituted to form
hydrogen bonds with the side chains of Ser5.42, Ser5.43 and Ser5.46 within transmembrane helix 5, which
are crucial for the activation of aminergic GPCRs by catecholamines and heterocyclic bioisosteres
thereof.15ꢀ18 Inspired by such Hꢀbond forming headgroups,19 we conceived a formal structural
hybridization of the 1,4ꢀDAP substructure of ariprazole and the heterocyclic neurotransmitter
surrogates present in the high affinity β2ꢀAR agonists indacaterol,20 procaterol,21, 22 olodaterol23 and BIꢀ
24
167107 (4),17,
leading to a novel chemotype of dopaminergics (Chart 2). Both, the 8ꢀ
hydroxyquinolinꢀ2(1H)ꢀone and the 5ꢀhydroxyꢀ2Hꢀbenzo[b][1,4]oxazinꢀ3(4H)ꢀone substructures have
been shown to mimic the phenolic hydroxyls of the endogenous β2ꢀAR agonists norepinephrine.17, 25
ACS Paragon Plus Environment

Page 5 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Besides these two heterocycles, we included the respective oxygenꢀanalog 8ꢀhydroxycoumarine, an 8ꢀ
hydroxyꢀsubstituted quinoline and a formally reduced 1,4ꢀbenzodioxane. Although benzodioxaneꢀ
substructures have been employed previously within heterocyclic 1,4ꢀDAPs,26 these ligands were
lacking the additional hydroxyl functionality. In addition to the variation of the aromatic head group,
we intended to replace the typical 1,4ꢀpiperazine by a homologous 1,4ꢀdiazepane (1,4ꢀ
homopiperazine), since this modification has been previously reported to foster D2R activation in a
putatively βꢀarrestinꢀbiased manner.14, 27 As lipophilic appendages, we either employed the 3,4ꢀ
dihydroquinolinꢀ2(1H)ꢀone substructure of aripiprazole, or used a pyrazolo[1,5ꢀa]pyridine, a moiety
that has been previously shown to convey high dopamine receptor affinity.28, 29 In both cases, we
connected the lipophilic moiety to the piperazine in analogy to the lead structure aripiprazole, using a
flexible butoxyꢀspacer.
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
Chart 2. Formal structural hybridization of the D2R partial agonist aripiprazole and the known β2ꢀAR
agonists procaterol and 4, leading to novel dopaminergics.
Synthesis. Synthesis of heteroarylpiperazines comprising a 3,4ꢀdihydroquinolinꢀ2(1H)ꢀone or
pyrazolo[1,5ꢀa]pyridine moiety and a butoxy spacer was established based on a convergent synthesis
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 68
1
2
3
4
5
6
7
8
9
strategy. In general, heteroarylpiperazines and their homopiperazine analogs were synthesized starting
from the unsaturated heterocycles, which were benzylꢀprotected and brominated to allow introduction
of Nꢀbocꢀprotected 1,4ꢀpiperazine or 1,4ꢀdiazepane building blocks by BuchwaldꢀHartwig amination
reactions (Scheme 1). These head groups were then reacted with the 4ꢀbromobutoxyꢀsubstituted 3,4ꢀ
dihydroquinolinꢀ2(1H)ꢀone or pyrazolo[1,5ꢀa]pyridine moieties to give the target compounds (Scheme
2). In detail, the synthesis of the 8ꢀ(benzyloxy)ꢀ5ꢀbromoquinolinꢀ2(1H)ꢀone building block 5c25, 30 was
performed starting from commercially available 8ꢀhydroxyquinolinꢀ2ꢀ(1H)ꢀone which was reacted with
benzyl bromide using DBU as a base followed by bromination of the aromatic core in the paraꢀposition
of the benzylꢀprotected hydroxy group yielding the intermediates 5b and 5c, respectively.
Unfortunately, reaction of 5c and bocꢀprotected piperazine was not successful. Therefore, a second Oꢀ
benzylation was performed, yielding the 2,8ꢀbis(benzyloxy)ꢀ5ꢀbromoquinoline 5d. The synthesis of 8ꢀ
hydroxycoumarineꢀbased ligands was established in analogy to described procedures,31, 32 starting from
2,3ꢀdihydroxybenzaldehyde, which was subjected to a Wittig reaction and intramolecular cyclization
with carbethoxycarbonylmethylene triphenylphosphorane in diethylaniline. The phenol 6a was
subsequently converted into the benzyl ether 6b and reaction with bromine in THF yielded the central
building block 8ꢀ(benzyloxy)ꢀ5ꢀbromoꢀ2Hꢀchromenꢀ2ꢀone (6c). The 5ꢀhydroxyꢀ2Hꢀ1,4ꢀbenzoxazinꢀ
3(4H)ꢀone 7a was prepared following a previously described three step procedure. First, 2ꢀ
nitroresorcinol was converted to 2ꢀaminoresorcinol by catalytic hydrogenation. The primary aromatic
amine was reacted with chloroacetyl chloride, before the final cyclization under basic conditions
afforded 7a.19 Subsequently, the phenol group was reacted with benzyl bromide using DBU as a base
and the aromatic core was brominated in the paraꢀposition of the benzylꢀprotected hydroxyl, giving
access to the intermediates 7b and 7c.19 The lactam nitrogen required protection by addition of a
second benzyl group, which was introduced by reaction with benzyl bromide, using elevated
temperatures and potassium carbonate as a base to afford the intermediate 7d. Synthesis of the 8ꢀ
hydroxybenzoꢀ1,4ꢀdioxane head group started from pyrogallol which was converted to 8ꢀ
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Page 7 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
hydroxybenzoꢀ1,4ꢀdioxane (8a) by reaction with 1,2ꢀdibromoethane following a previously described
protocol.33 Subsequently, the phenolic hydroxyl group was benzylꢀprotected (8b) and a bromine
residue was introduced into the paraꢀposition of the benzylꢀprotected hydroxyl, yielding intermediate
8c. Benzyl protection34 followed by bromination of commercially available 8ꢀhydroxyquinoline
afforded 9c,35 which served as precursor for 8ꢀhydroxyquinolineꢀderived ligands. The five different
bromoꢀsubstituted heterocycles were subsequently reacted with bocꢀprotected 1,4ꢀpiperazine in
BuchwaldꢀHartwig crossꢀcoupling reactions, using conditions described for the reaction of 9c with
piperidine.35 All reactions were carried out in toluene using either Cs2CO3 or NaOtBu as a base. As
palladium species, Pd(OAc)2 or Pd2(dba)3 together with racemic BINAP were used. Using the reaction
sequence elaborated for the preparation of the arylpiperazines 5e, 6d, 7e, 8d and 9d, reaction of the
bromoꢀsubstituted precursors 5d, 6c and 7d with bocꢀprotected 1,4ꢀdiazepane yielded the
homopiperazine analogs 5f, 6e and 7f. Finally, cleavage of the tertꢀbutyloxycarbonyl groups with TFA
in dichloromethane resulted in formation of five 1,4ꢀheteroarylpiperazines (5g, 6f, 7g, 8e, 9e) and three
1,4ꢀheteroarylhomopiperazines (5h, 6g, 7h). For the synthesis of the final compounds, the secondary
amines of the heteroarylpiperazines (5g, 6f, 7g, 8e, 9e) and ꢀhomopiperazines (5h, 6g, 7h) were
subjected to nucleophilic displacement reactions with 7ꢀ(4ꢀbromobutoxy)ꢀ3,4ꢀdihydroquinolinꢀ2(1H)ꢀ
one36 or 5ꢀ(4ꢀbromobutoxy)ꢀpyrazolo[1,5ꢀa]pyridine28, 37 (Scheme 2). In both cases, the alkylbromides
were activated by refluxing with sodium iodide.28, 36 Finally, the remaining benzyl protecting groups
were cleaved by hydrogenation or under acidic conditions, affording the heteroarylpiperazines 12aꢀe,
16aꢀe and heteroarylhomopiperazines 13aꢀc and 17a.
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 addition to the collection of ligands bearing a secondary moiety attached via a butoxyꢀspacer,
we intended to complement our investigations with a different lipophilic appendage (Scheme 3). Very
recently, we have developed a vanillinꢀderived scaffold linked to an amidopropyl or aminobutyl spacer
as a useful moiety for specific molecular tools,38, 39 and functionally selective D2R ligands.30 Thus, we
ligated the homopiperazineꢀ8ꢀhydroxyquinolinꢀ2(1H)ꢀone 5h as a representative head group with the
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 68
1
2
3
carbaldehyde function of the vanillinꢀtype appendages by reductive amination.30, 38 The intermediates
4
5
18a,b were converted to the final compounds 19a,b by cleavage of the benzyl protection groups with
6
7
methanesulfonic
acid
in
toluene.
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
ACS Paragon Plus Environment

Page 9 of 68
Journal of Medicinal Chemistry
Scheme 1.a Synthesis of heterocyclic aromatic piperazine and homopiperazine head groups.
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
aReagents and conditions: (a) DBU, benzyl bromide, isopropyl alcohol, rt, 1.5 h (70ꢀ88 %); (b) Br2, AcOH, rt, 4 h (53ꢀ84 %);
(c) K2CO3, benzyl bromide, DMF, 50°C, 24 h or rt, 2 h (57ꢀ91 %); (d) Pd2(dba)3 or Pd(OAc)2, racꢀBINAP, NaOtBu, 1ꢀbocꢀ
piperazine or 1ꢀbocꢀhomopiperazine, toluene, 85°C, 115°C or reflux, 24ꢀ48 h (63ꢀ73 %); (e) TFA, CH2Cl2, rt, 2 h (51ꢀ97 %);
(f) Br2, THF, 0°C, 2 h, then rt, 16 h (47 %); (g) Pd2(dba)3 or Pd(OAc)2, racꢀBINAP, Cs2CO3, 1ꢀbocꢀpiperazine or 1ꢀbocꢀ
homopiperazine, toluene, 85°C or reflux, 24ꢀ72
h
(38ꢀ66 %); (h) Br2, CH2Cl2, rt,
5
h
(55ꢀ75 %).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 68
1
2
3
4
Scheme 2.a Synthesis of full length dopaminergics.
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
aReagents and conditions: (a) 7ꢀ(4ꢀbromobutoxy)ꢀ3,4ꢀdihydroquinolinꢀ2(1H)ꢀone, NaI, acetonitrile, reflux 30
min, then addition of triethylamine in acetonitrile, 80°C, 16h (45ꢀ74 %); (b) 5ꢀ(4ꢀbromobutoxy)ꢀpyrazolo[1,5ꢀ
a]pyridine, NaI, acetonitrile, reflux 30 min, then addition of triethylamine in acetonitrile, 80°C, 16h (65ꢀ77 %);
(c) 10% Pd/C in acetic acid/methanol, rt, 4 h (63ꢀ79 %) or TFA in toluene, reflux, 8 h (57ꢀ79 %) or MsOH in
toluene, reflux, 3 h (48ꢀ59 %).
Scheme 3.a Synthesis of heteroarylpiperazines with alternate appendages.
aReagents
methoxybenzamide
and
conditions:
(a)
4ꢀ(3ꢀ(1ꢀbutylꢀ1Hꢀ1,2,3ꢀtriazolꢀ4ꢀyl)propoxy)ꢀNꢀ(3ꢀoxopropyl)ꢀ3ꢀ
or 4ꢀ(3ꢀ(1ꢀbutylꢀ1Hꢀ1,2,3ꢀtriazolꢀ4ꢀyl)propoxy)ꢀNꢀ(4ꢀoxobutyl)ꢀ3ꢀmethoxybenzamide,
NaBH(OAc)3, dichloroethane, rt, 6 h (80ꢀ89 %); (b) MsOH in toluene, reflux, 3 h (43ꢀ74 %).
ACS Paragon Plus Environment

Page 11 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Binding affinities. In order to determine the impact of the structural modifications on dopamine
receptor recognition, we performed radioligand displacement studies with the dopamine receptor
subtypes D1R, D2LR, D2SR, D3R, D4R and D5R. The resulting affinity and selectivity profiles were
compared to those of the chemically similar reference ligands aripiprazole and 1 (Table 1). Binding
affinities were obtained by measuring the ability of the target compounds to compete with the
radioligands [3H]SCH23990 or [3H]spiperone and membranes from cells expressing the respective
receptor subtypes. In general, the series of target compounds showed only weak affinity for the D1R
and D5R subtypes, while they bound to the D2Rꢀlike family with low nanomolar to even subnanomolar
Ki values. Indeed, heteroarylpiperazines comprising a 3,4ꢀdihydroquinolinꢀ2(1H)ꢀone appendage (12aꢀ
e) were highly similar to the reference agents and displayed binding affinities ranging from 0.87ꢀ4.8,
0.67ꢀ2.2 and 1.0ꢀ12 nM at D2LR, D2SR and D3R, respectively. Interestingly, the incorporation of a
methylene unit into the 6ꢀmembered cyclic diamine (piperazine vs. homopiperazine) did not
significantly alter the binding affinities for ligands with 8ꢀhydroxyquinolinone (13a) and 5ꢀ
hydroxybenzoxazinone (13c) head groups (Ki 0.81ꢀ2.0 nM for D2S/LR and D3R), but induced a 15ꢀ to
40ꢀfold loss of affinity for the homopiperazine 13b (Ki 37ꢀ240 nM) comprising an 8ꢀhydroxycoumarine
head group. When the 3,4ꢀdihydroquinolinꢀ2(1H)ꢀone appendage was replaced by a pyrazolo[1,5ꢀ
a]pyridine moiety (16aꢀe), binding affinities for D2LR and D2SR were decreased. Interestingly, the
extent of this effect was found to depend on the employed heteroaryl pharmacophore. While the
quinolinone 16a and the benzoxazine 16c still showed high affinity towards D2LR and D2SR (Ki 0.86ꢀ
8.3 nM), derivatives lacking an the endocyclic hydrogen bond donor at the heterocyclic head group
showed substantially lower binding affinities for D2LR/D2SR (Ki 19ꢀ110 nM). The homopiperazine 17a
showed lower D2R binding affinity (Ki 78 and 46 nM for D2LR and D2SR, respectively). Notably,
affinities for the closely related D3R subtype did not follow the same trend, as the pyrazolo[1,5ꢀ
a]pyridine derivatives 16aꢀc and 17a displayed low nanomolar or even subꢀnanomolar binding
affinities (Ki 0.71ꢀ8.5 nM) that were highly similar to those of the reference agents aripiprazole and 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
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 68
1
2
3
(Ki 1.9 and 2.9 nM). At the D4R subtype, only the heteroarylpiperazines 12c and 16c sharing the
4
5
benzoxazinone head group displayed low nanomolar affinity (8.3 and 4.8 nM).
6
7
8
9
Additional radioligand binding studies were performed to determine the compounds’ affinities
towards related serotonergic (5ꢀHT1AR, 5ꢀHT2AR) and adrenergic (α1ꢀAR) receptors using the
radioligands [3H]WAY100635, [3H]ketanserin or [3H]prazosin. In comparison to the reference agents
aripiprazole and 1, our newly developed compounds showed substantially altered binding properties.
At the 5ꢀHT1AR, the test compounds’ affinities ranged from 12 nM to 1,200 nM, with the highest
binding affinities observed for 12d, 12e and 16d containing an 8ꢀhydroxybenzoꢀ1,4ꢀdioxane or an 8ꢀ
hydroxyquinoline head group. 5ꢀHT2AR affinities were substantially lower than for aripiprazole or 1,
with the highest affinities observed for the 8ꢀhydroxyquinolines 12e and 16e and the 8ꢀhydroxybenzoꢀ
1,4ꢀdioxane 12d. Thus the vast majority of novel target compounds displayed high selectivity for D2R
over 5ꢀHT2AR, while maintaining moderate affinity for the 5ꢀHT1AR, a receptor that has been identified
as additional target in the treatment of PD.40 Compared to the typical 1,4ꢀDAP structures, inclusion of a
propylamidoꢀ or butyramidoꢀvanillinꢀtype appendage (19a,b) led to decreased binding affinities at
D2LR, D2SR and D4R. Notably, 19a,b displayed oneꢀdigit nanomolar Ki values at D3R (1.6 and 3.2 nM)
leading to an approximately 100ꢀfold selectivity for D3R over D2R.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Page 13 of 68
Journal of Medicinal Chemistry
Table 1. Binding affinities of the heteroarylpiperazines towards dopaminergic, serotonergic and αꢀadrenergic
receptors.
1
2
Ki ± SD [nM]a
3
4
5
6
7
8
comp.
aripip.
R1
n
1
R2
hD1Rb
hD2LRc
hD2SRc
hD3Rc
hD4Rc
hD5Rb p5ꢀHT1ARd h5ꢀHT2ARe
pα1Rf
310
± 140
0.52
± 0.17
0.38
± 0.16
2.9
± 0.66
63
± 17
1600
88
3.6
17
± 1
± 500
± 7
± 1.1
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
340
± 40
0.53
± 0.37
0.20
± 0.06
1.9
± 0.52
76
± 22
1,100
± 140
8.7
± 4.8
2.0
± 0.4
16
± 3
2
1
1
1
1
1
2
2
2
1
1
1
1
1
2
1
5,400
± 1,700
0.89
± 0.71
1.0
± 0.8
12
± 7
66
9,300
48
± 11
2,200
± 500
96
± 11
12a
12b
12c
12d
12e
13a
13b
13c
16a
16b
16c
16d
16e
17a
19a
± 20 ± 1,600
4.8
± 3.6
2.0
± 1.0
7.7
± 3.9
870
230
± 190
1,400
± 700
>20,000
>20,000
± 380
>20,000
12,000
± 3,200
0.87
± 0.81
0.67
± 0.24
1.0
± 0.8
8.3
4,200
15
± 1
430
± 60
20
± 2
± 2.9 ± 2,100
4,500
± 700
2.9
± 0.6
2.2
± 0.4
38
± 10
300
16,000
12
± 11
160
± 40
130
± 60
± 70 ± 5,000
1,500
± 1,000
3.4
± 1.5
1.5
± 0.4
27
± 16
94
± 64
3,400
± 920
12
± 4
39
± 24
32
± 7
11,000
± 8,000
1.6
± 0.7
0.81
± 0.48
1.9
± 0.4
75
± 37
150
± 60
3,400
± 1,600
61
± 25
>20,000
100
± 40
37
± 20
240
± 90
5,000
± 100
800
± 430
2,300
± 100
900
± 60
>20,000
>20,000
15,000
2,700
± 2,600
1.6
± 1.1
1.6
± 0.9
2.0
± 0.9
60
280
± 170
2,000
± 1,600
200
± 80
± 30 ± 12,000
4,700
± 4,200
8.3
± 2.7
3.7
± 0.4
3.8
± 1.3
31 13,000
± 10 ± 9,000
63
± 25
3,600
± 3,500
87
± 23
40
± 38
19
± 6
8.5
± 5.1
270
17
± 4
2,800
± 600
160
± 10
>20,000
>20,000
± 70
2,300
± 300
2.0
± 1.4
0.86
± 0.44
0.71
± 0.57
4.8
16,000
25
± 4
1,300
± 1,200
15
± 1
± 3.0 ± 15,000
5,300
± 5,200
110
± 30
68
± 13
40
± 14
160
19,000
13
± 0
480
± 420
110
± 20
± 50 ± 1,000
1,100
± 500
33
± 7
25
± 6
50
± 10
55
5,800
28
± 4
200
± 20
93
± 5
± 6 ± 3,500
13,000
± 2,000
78
± 52
46
± 17
3.4
± 2.5
190
1,200
± 400
9,000
± 4,300
430
± 40
>20,000
± 70
17,000
± 14,000
350
± 170
440
± 250
1.6
± 1.0
3,200
3,300
± 500
4,200
± 700
1,000
± 150
>20,000
± 600
4,000
± 3,100
200
± 120
220
± 160
3.2
± 1.6
2,100
1,400
± 600
1,500
± 1,000
810
± 50
>20,000
± 500
19b
a
b 3
Data represent mean ± S.D. of two to fifteen independent experiments, each performed in triplicate. Radioligands: [ H]SCH23390,
c[3H]spiperone, d[3H]WAY100635, e[3H]ketanserin, f[3H]prazosin.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 68
1
2
3
4
5
6
7
8
9
Functional activity at D2R. The influence of the individual structural modifications on the
ligands’ ability to activate D2R was evaluated in a βꢀarrestinꢀ2 recruitment assay. The arrestin receptor
interaction was initially discovered to attenuate or terminate G proteinꢀmediated signaling.41 There is
now broad consensus that βꢀarrestins also represent signal transducers on their own.42 D2R activation
mediated through βꢀarrestinꢀ2 leads to the G proteinꢀindependent activation of GSK3β, an enzyme
which has been implicated in hyperdopaminergic pathophysiological conditions of neuropsychiatric
disorders such as schizophrenia.43 In contrast, studies with animal models of PD suggest that enhanced
stimulation of βꢀarrestinꢀ2 signaling may have beneficial effects on the control of locomotion while
reducing dyskinesia frequently occurring in PD treatment.12 We measured the recruitment of βꢀarrestinꢀ
2 to ligandꢀactivated D2SR employing the DiscoverX PathHunter technology in HEK293 cells as
described previously,28 and compared the results of our target compounds to those of known D2R
agonists and the 1,4ꢀDAPs aripiprazole and 1 (Table 2). When the reference agonist quinpirole, the
endogenous agonist dopamine and the antiparkinsonian drugs rotigotine, pramipexole and ropinirole
were examined for their ability to induce βꢀarrestinꢀ2 recruitment, sigmoid dose response curves were
observed in all cases (Supplementary Figure S1). The ligands displayed similar efficacies ranging from
full agonist behavior to very strong partial agonism observed for ropinirole (Emax 83 %). The highest
potency was observed for rotigotine (EC50 4.1 nM), while about 100ꢀfold higher concentrations were
necessary to activate D2SR with the endogenous agonist dopamine (EC50 390 nM). In comparison to
these relatively similar effects of the reference ligands, the novel heteroarylpiperazines and their
homopiperazine analogs sampled a wide range of efficacies (Supplementary Figure S2). Indeed, the
maximum efficacy was found to depend on two main factors: the nature of the orthosteric heteroaryl
moiety and the ring size of the central diamine. More specifically, piperazines 12b,d,e and 16b,d,e
comprising an 8ꢀhydroxybenzoꢀ1,4ꢀdioxane, an 8ꢀhydroxyquinoline or an 8ꢀhydroxycoumarine head
group were devoid of agonistic properties in the βꢀarrestinꢀ2 recruitment assay, similar to the results
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Page 15 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
obtained for the antipsychotic drug aripiprazole. In contrast, ligands 12a,c and 16a,c sharing an 5ꢀ
hydroxybenzoxazinꢀ3ꢀone or 8ꢀhydroxyquinolinꢀ2(1H)ꢀone substructure showed partial agonist
activities with Emax values between 21 and 44 % and potencies comparable to quinpirole and dopamine
(EC50 62ꢀ240 nM). Notably, enlargement of the cyclic diamine moiety from a 6ꢀmembered (piperazine)
to a 7ꢀmembered ring (homopiperazine) increased the intrinsic efficacy of the compounds. Thus, the βꢀ
arrestinꢀ2 recruitment assay revealed an Emax value of 75% for the homopiperazine 13c comprising a 5ꢀ
hydroxybenzoxazinꢀ3ꢀone head group. Moreover, the 8ꢀhydroxyquinolinꢀ2(1H)ꢀoneꢀhomopiperazines
13a and 17a behaved as full agonists (Emax 90ꢀ102 %, Figure 1a,c). These findings are in accordance
with earlier reports on an increase in intrinsic efficacy for the diazepane/piperazine replacement in
classical 1,4ꢀDAP scaffolds.27 However, the presence of the homopiperazine subunit does not lead to
pronounced recruitment of βꢀarrestinꢀ2 on its own, since the reference ligand 1 only induced marginal
recruitment of βꢀarrestinꢀ2 under these assay conditions (Supplementary Figure S1). The main
difference between 1, its closely related analogs 2 and 3 (Chart 1)14, 27 and our newly synthesized
ligands can be found in the heterocyclic moiety, were these ligands lack functional groups capable to
interact with the polar serine residues in the orthosteric binding pocket of the receptor. Similar to the
rankꢀorder in efficacies observed among the piperazines 12aꢀe, the 8ꢀhydroxycoumarineꢀ
homopiperazine 13b was less effective (Emax 21 %, Figure 1b) than 13a or 13c. These findings
underline the importance of the presence of an additional hydrogen bond donor rather than an Hꢀbond
acceptor within the aromatic heterocycle, since 13b and results from 13a in a formal NH to O
exchange. These results are in excellent agreement with the binding pose of the highꢀaffinity agonist 4
within the activeꢀstate structure of the β2ꢀAR.17 Moreover, studies on the β1ꢀAR in complex with
different agonists18 and the β2ꢀAR crystal structure in complex with a covalent agonist25 highlighted the
importance of the hydrogen bond network stabilized by catecholamine surrogates for receptor
activation. All homopiperazines investigated (13aꢀc, 17a) displayed potencies similar to the
endogenous agonist dopamine (EC50 190ꢀ470 nM). In agreement with previous results pointing towards
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 68
1
2
3
4
5
6
7
8
a profitable effect of the vanillinꢀtype appendage for the recruitment of βꢀarrestinꢀ2,30 incorporation of
this lipophilic moiety in combination with the 8ꢀhydroxyquinolinꢀ2(1H)ꢀoneꢀsubstituted
homopiperazine head group led to pronounced partial agonist properties for the recruitment of βꢀ
arrestinꢀ2 at D2SR. However, in accordance with the reduced D2R affinities, ligands 19a,b displayed
relatively weak potencies (EC50 680 and 1,600 nM).
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 2. Ligandꢀinduced recruitment of βꢀarrestinꢀ2 at D2SR.
βꢀarrrestinꢀ2 recruitment at D2SRa
EC50 ± SEM
[nM]
Emax ± SEM
[%]b
compound
quinpirole
dopamine
rotigotine
ropinirole
pramipexole
aripiprazole
1
52 ± 6
390 ± 20
4.1 ± 1.0
89 ± 5
100 ± 1
96 ± 2
101 ± 4
83 ± 4
90 ± 6
13 ± 3
≤ 10
27 ± 4
1100 ± 600
n.d.c
93 ± 44
n.d.c
31 ± 4
≤ 10
12a
12b
92 ± 36
n.d.c
44 ± 4
≤ 10
12c
12d
n.d.c
≤ 10
12e
190 ± 40
220 ± 30
470 ± 80
240 ± 60
n.d.c
102 ± 4
21 ± 2
75 ± 2
21 ± 1
≤ 10
13a
13b
13c
16a
16b
62 ± 9
43 ± 3
≤ 10
16c
n.d.c
16d
n.d.c
≤ 10
16e
370 ± 100
680 ± 70
1600 ± 600
90 ± 3
90 ± 2
76 ± 5
17a
19a
19b
ACS Paragon Plus Environment

Page 17 of 68
Journal of Medicinal Chemistry
1
2
3
a
b
Data represent mean ± S.E.M. of four to ten independent experiments, each performed in duplicate. Emax relative to the
effect of the reference agonist quinpirole. cn.d., not determined
4
5
6
7
8
9
In an attempt to identify βꢀarrestinꢀbiased D2R agonists, a subset of the target compounds was
investigated for their intrinsic activity towards the activation of canonical signaling at D2SR. Thus, we
studied ligandꢀpromoted nucleotide exchange in membrane preparations from HEK293T cells coꢀ
expressing D2SR and a pertussis toxin insensitive variant of the Gαo1 subunit44 as a proximal measure of
G protein activation.45 Although D2R is known to promiscuously couple to all members of the Gαi/o
family,46 Gαo proteins have been identified as predominant mediators of D2Rꢀsignaling in the central
nervous system (CNS).47 Incubation of the membranes with the endogenous agonist dopamine and the
antiparkinsonian drugs rotigotine, ropinirole and pramipexole led to a similar degree of nucleotide
exchange compared to the reference agonist quinpirole (Table 3). While quinpirole, dopamine,
pramipexole and ropinirole showed relatively similar potencies (EC50 100ꢀ330 nM), rotigotine was
effective at 30ꢀ to 100ꢀfold lower concentration (EC50 3.2 nM). These results are in good agreement
with the recruitment of βꢀarrestinꢀ2, when rotigotine was also found to be the most potent agonist (EC50
4.1 nM). Relative to these high intrinsic efficacies, aripiprazole and 1 were relatively weak partial
agonists (Emax 14ꢀ17 %). In particular, very high concentrations of 1 were necessary to elicit guanine
nucleotide exchange at D2SR. These findings are in good agreement with literature reports indicating a
low tendency of 1 to inhibit cAMP formation.14 Compared to 2,3ꢀdichlorophenylꢀsubstituted
piperazines or homopiperazines, hydroxylated heteroarylhomopiperazines 13aꢀc and 17a showed
substantially higher intrinsic efficacies for the induction of canonical D2SR signaling (Emax 26ꢀ55 %).
Potencies of 13a, 13c and 17a were found to be slightly higher than those of dopamine or quinpirole
(EC50 67ꢀ90 nM), while 13b only elicited significant G protein activation at concentrations above 1 µM
(Figure 1b). The target compounds 13a,c and 17a were about 50 % less efficacious in promoting
nucleotide exchange compared to the recruitment of βꢀarrestinꢀ2 (Emax 26ꢀ55 % and 75ꢀ102 %,
respectively). Although relative efficacies suggest a preferential engagement βꢀarrestins over G
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 68
1
2
3
4
5
6
7
8
9
proteins, a quantitative bias analysis employing the operational model of agonism did not reveal
significantly biased agonists among these target compounds, most likely due to the increased potencies
for G protein activation. In contrast, the lactone 13b, renouncing an endocyclic hydrogen bond donor at
the heterocyclic head group, was identified as a βꢀarrestinꢀbiased ligand, when the endogenous agonist
dopamine was used as a reference (Supplementary Figure S3a). Bias was also detected for the very
weak partial agonist 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
Table 3. Ligandꢀinduced nucleotide exchange determined with membranes from HEK293T cells
coexpressing D2SR and Gαo1 and βꢀarrestinꢀ2 recruitment promoted by D2SR in presence of GRK2.
[35S]GTPγS
(D2SR + Gαo1
βꢀarrrestinꢀ2 recruitment
(D2SR + GRK2)a
bias factord
(D2SR)
ꢄ
a
)
EC50 ± SEM Emax ± SEM EC50 ± SEM Emax ± SEM
∆∆ꢀꢁꢂ ꢃ
ꢇ
compound
[nM]
[%]b
[nM]
[%]b
ꢅꢆ
quinpirole
dopamine
rotigotine
ropinirole
270 ± 30
100 ± 1
16 ± 1
100 ± 1
1.00 ± 0.07e
0.00 ± 0.08
0.15 ± 0.10
0.99 ± 0.08e
0.79 ± 0.09
ꢀ0.14 ± 0.35
2.42 ± 0.48e
ꢀ0.48 ± 0.10
1.91 ± 0.19e
ꢀ0.58 ± 0.20
ꢀ0.19 ± 0.10
200 ± 20
3.2 ± 0.5
330 ± 60
93 ± 2
91 ± 2
92 ± 2
102 ± 3
14 ± 2
17 ± 7c
55 ± 4
36 ± 6c
26 ± 2
56 ± 4
160 ± 20
4.4 ± 1.1
30 ± 6
97 ± 2
106 ± 4
104 ± 3
100 ± 3
43 ± 5
pramipexole 100 ± 30
aripiprazole 130 ± 60
9.6 ± 1.4
740 ± 180
270 ± 80
200 ± 20
130 ± 20
630 ± 110
250 ± 40
> 5 µMc
19 ± 3
1
67 ± 20
> 5µMc
78 ± 25
90 ± 16
110 ± 6
51 ± 5
13a
13b
13c
17a
94 ± 5
103 ± 2
aData represent mean ± S.E.M. of five to fifteen independent experiments, each performed in triplicate. bEmax relative to the
effect of the reference agonist quinpirole. cNo complete dose response curves obtained, Emax determined at 30 µM. dLigand
bias quantified by the operational model, values < 0 indicate preference for GTP binding, values > 0 preferential
e
recruitment of βꢀarrestinꢀ2. Significant bias determined by oneꢀway ANOVA, followed by Dunnet’s post hoc test, p <
0.05.
Since it is known that recruitment of βꢀarrestins is greatly facilitated upon adequate receptor
phosphorylation,48 we performed additional βꢀarrestinꢀ2 recruitment assays in the presence of GRK2, a
kinase that has been reported to be relevant for the interaction of βꢀarrestinꢀ2 and D2R.26, 49, 50 Dose
response curves of the endogenous agonist dopamine and the reference agonists quinpirole, rotigotine,
ACS Paragon Plus Environment

Page 19 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
ropinirole and pramipexole revealed highly similar maximum efficacies (Emax 97ꢀ104 %) under these
conditions. While the potency of rotigotine was unaffected by the presence of GRK2 (EC50 4.4 nM), a
2ꢀ to 3ꢀfold increase was observed for the four other agonists including the endogenous
neurotransmitter dopamine (EC50 9.6ꢀ160 nM). In good agreement with earlier reports, an increase in
efficacy was observed for the weak partial agonists aripiprazole and 1 (Emax 43 and 19 %, respectively).
Among our newly developed hydroxylated heterocyclic homopiperazines, coexpression of GRK2
substantially increased the efficacy of 13b (Emax 51 %, Figure 1b) and 13c (Emax 94 %), comprising an
8ꢀhydroxycoumarine or a 5ꢀhydroxybenzoxazinꢀ3ꢀone head group, respectively. For the 8ꢀ
hydroxyquinolinꢀ2(1H)ꢀones 13a and 17a only minor changes in efficacy or potency were observed
(Figure 1a,c), probably indicating a minor contribution of GRK2 to the signaling panorama of these
two ligands. A quantitative comparison of the signaling profiles confirmed the preference for βꢀ
arrestinꢀ2 recruitment for the 8ꢀhydroxycoumarine 13b and the reference agent 1 (ꢁꢁlog(τ/KA) 1.91
and 2.42, Figure 1d). No such bias was observed for the ligands 13a,c and 17a, although the signaling
profiles of these two ligands are substantially different from those of typical antiparkinsonian drugs
(e.g. rotigotine) or the endogenous agonist dopamine (Supplementary Figures S1ꢀ3). These results are
most likely caused by the detrimental changes of potency and efficacy observed for 13a,c and 17a
between the different test systems, and the relative importance attributed to these influencing factors in
the operational model of agonism. In order to relate these distinct signaling profiles to pharmacological
outcomes, and thus to estimate the therapeutic value of such compounds future studies, potentially also
involving βꢀarrestin knockꢀout animals, will be required.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 68
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. Doseꢀresponse curves at D2SR and metabolic stability of target compounds 13a,b and 17a.
(aꢀc) Doseꢀresponse curves for D2SRꢀmediated activation of Gαo1 (grey) and the recruitment of βꢀ
arrestinꢀ2 in absence (blue) or presence (orange) of GRK2. (d) Bias factors for D2SR activation were
calculated using the operational model of agonism. Positive values for ꢁꢁlog(τ/KA) indicate a
preference for βꢀarrestinꢀ2 recruitment. (e) Metabolic stability of 13a and 17a assessed in rat liver
microsomes: distribution between metabolized substrate, substrate nonspecifically bound to matrix and
the remaining fraction after 60 min of incubation. (f) Inhibition of hERG tail currents by 13a was
measured to estimate susceptibility of cardiotoxicity. Data represent mean ± SEM of at least three
independent experiments.
Metabolism and inhibition of hERG tail currents. In order to assess the drugꢀlike properties
of our target compounds we calculated octanol/water partition coefficients as described previously
(Supplementary Table S1).28 Compared to the typical 1,4ꢀDAP aripiprazole, hydroxyꢀsubstituted
heteroarylpiperazines and their homopiperazine analogs were predicted to be less lipophilic. With the
exception of the most hydrophilic compounds 12c, 13c and 16c sharing the 5ꢀhydroxybenzoxazinone
head group, the ligands’ lipophilicity was predicted to be similar to the range covered by the approved
D2R agonists ropinirole, pramipexole and rotigotine. Furthermore, with the exception of 19a,b, none of
the compounds hit constraints predicting poor solubility or permeability as indicated by Lipinski’s rule
of five.51 Metal chelating properties for various divalent cations are well known characteristics of 8ꢀ
ACS Paragon Plus Environment

Page 21 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
hydroxyquinoline substructures and have been associated to undesired side effects.52 Although
chelating properties are likely for some of the presented ligands, no toxic effects were observed under
the conditions of our functional experiments.
6
7
8
9
Besides adequate physicochemical properties, sufficient metabolic stability is a general
requirement in drug development. Thus, to determine the metabolic stability of the novel
heteroarylpiperazines, we explored the representative ligands 13a and 17a for their propensity to
undergo phase I transformations using rat liver microsomes in comparison to the antiparkinsonian drug
rotigotine. In agreement with previous findings,53 rotigotine was extensively metabolized within 15
minutes of incubation, and a number of known metabolites were identified using mass spectrometry
(Supplementary Figures S4, S5). The half life and intrinsic clearance for rotigotine were determined to
be 5.8 min and 240 µL/min/mg protein, respectively (Supplementary Figure S4).54, 55 In contrast, only
minor metabolization of the pyrazolo[1,5ꢀa]pyridine 17a was observed (T1/2 51 min, CLint 27
µL/min/mg protein), while the 3,4ꢀdihydroquinolinꢀ2(1H)ꢀone 13a was hardly metabolized at all (T1/2
130 min, CLint 11 µL/min/mg protein). Based on the intrinsic clearance, 17a can be regarded as
intermediately metabolized, while 13a falls into the low clearance category.54, 55 Both ligands revealed
binding to matrix protein (~27 %) in control experiments when the enzymatic transformation was
prevented by the absence of the cofactor NADPH (Figure 1e, Supplementary Figures S4, S5).
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
QTcꢀinterval prolongation leading to the very rare but potentially lethal Torsades de Pointes
(TdP) arrhythmia is a serious complication associated with pharmacotherapy.56, 57 Although not the
only cause, inhibition of the delayed rectifier potassium IKr current mediated by hERG (humanꢀetherꢀàꢀ
goꢀgo related gene) channels is predictive for QTc prolongation and the liability of a drugꢀlike
molecule to cause TdP.58 Drugꢀinduced arrhythmia has been observed for a number of dopaminergic
drugs, including the phenylpiperidine haloperidol and the aromatic piperazine ziprasidone.59 In order to
assess the susceptibility to induce arrhythmia, and hence estimate cardiotoxicity, the influence of a
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 68
1
2
3
4
5
6
7
8
9
representative hydroxyꢀsubstituted heterocyclic phenylpiperazine (13a) on the inhibition of hERG tail
currents was examined in vitro. Automated patchꢀclamp showed that no effect on hERG tail currents
was observed at concentrations up to 1 µM (Figure 1f). At higher concentration a slight inhibition was
observed (10 µM, 71.7 % current remaining), which was doseꢀdependently increased (50 µM, 21.5%
remaining). However, the inhibitory potency (IC50) of 13a was determined to be 19.7 µM, a
concentration that is 100ꢀfold higher than the EC50 required for βꢀarrestinꢀ2 recruitment at D2SR and
even 24,000ꢀfold higher than the binding affinity of 13a at D2SR (Ki 0.81 nM).
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
CONCLUSIONS
Applied in combination with a peripheral decarboxylase inhibitor, LꢀDOPA remains the ‘gold
standard’ in antiparkinsonian drug therapy. However, its use is limited by wearingꢀoff and the
occurrence of LꢀDOPA induced dyskinesia (LID) over the course of time.3 Very recently, it has been
demonstrated that dopamine receptorꢀmediated activation of βꢀarrestinꢀ2 is required for the beneficial
locomotor effects of LꢀDOPA and not involved in the pathogenesis of LIDs. Selective targeting of this
signal transducer, e.g. by biased ligands, may therefore represent a valuable novel concept in the
treatment of PD.12 Aiming to develop novel dopaminergics that activate βꢀarrestinꢀ2 in a functionally
selective manner via the interaction with D2R, we performed a formal structural hybridization of
typical 1,4ꢀDAPs and heterocyclic catecholamine surrogates leading to hydroxyꢀsubstituted heteroarylꢀ
1,4ꢀdisubstituted piperazines and homopiperazines. In contrast to previously described βꢀarrestinꢀ2
biased ligands for D2R,14, 27 these compounds are designed to mimic the interactions of the catechol
substructure of the endogenous ligand dopamine with crucial serine residues in the orthosteric binding
pocket. We found that two major factors determined the intrinsic efficacy of the ligands with this novel
scaffold – the presence of an additional endocyclic Hꢀbond donor rather than an Hꢀbond acceptor, and
the ringꢀsize of the cyclic diamine. Notably, ligands 13a,c and 17a comprising an 8ꢀhydroxyquinolinꢀ
ACS Paragon Plus Environment

Page 23 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2(1H)ꢀone or a 5ꢀhydroxyꢀ2Hꢀbenzo[b][1,4]oxazinꢀ3(4H)ꢀone heterocycle and a homopiperazine
substructure showed full agonist properties for the recruitment of βꢀarrestinꢀ2, while they were
substantially less efficacious for the induction of D2SRꢀpromoted guanine nucleotide exchange as a
measure of canonical Gαo protein activation. Although the 8ꢀhydroxycoumarineꢀsubstituted
homopiperazine 13b induced submaximal activation of both signaling pathways, this compound was
identified as highly βꢀarrestinꢀbiased using the operational model of agonism. A further evaluation of
representative homopiperazines 13a and 17a comprising a 3,4ꢀdihydroquinolinꢀ2(1H)ꢀone or
pyrazolo[1,5ꢀa]pyridine appendage revealed a low tendency to undergo metabolic transformation
reactions in rat liver microsomes and identified a minor liability to inhibit hERG channels, an important
offꢀtarget in the course of drugꢀdevelopment. The combination of typical agonistꢀtype elements from
known β2ꢀAR agonists such as procaterol21, 22 and 417, 24 with 1,4ꢀDAP substructures known for their
weak partial agonist characteristics60 proved successful for the design of novel βꢀarrestinꢀbiased D2R
agonists with unprecedented intrinsic efficacy. This novel scaffold might serve as a starting point for
the development of functionally selective PD therapeutics with a lower propensity to induce LID.
Nevertheless, the contribution of other dopamine receptor subtypes, especially the D1R, to the efficacy
and side effects associated with LꢀDOPA therapy should also be addressed in drug discovery.6, 12
Medicinal chemistry specifically targeting D1Rꢀpromoted βꢀarrestinꢀ2 recruitment may lead to further
promising innovations.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
EXPERIMENTAL PART
Chemistry. Reagents and dry solvents were of commercial quality and used as purchased. If
not stated otherwise, reactions were carried out under nitrogen atmosphere. MS was run on a BRUKER
ESQUIRE 2000 or on a BRUKER amaZon SL mass spectrometer using ESI ionization. HRMS was
run on a AB Sciex Triple TOF660 SCiex, source type ESI or at the Chair of Organic Chemistry,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 68
1
2
3
4
5
6
7
8
9
Friedrich Alexander University ErlangenꢀNuernberg on a Bruker Daltonik micrOTOF II focus or
Bruker Daltonik maXis 4G, source type ESI or APPI. NMR spectra were recorded on a Bruker Avance
400, Bruker Avance 360 or a Bruker Avance 600 spectrometer at 300 K in the solvents indicated.
Chemical shifts are given in ppm (δ) relative to TMS. IR spectra were obtained on a Jasco FT/IR 4100
spectrometer using a substance film on a NaCl crystal. Purification by column chromatography was
performed with silica gel 60. TLC analyses were performed using Merck 60 F254 aluminum plates in
combination with UV detection (254 nm) or ninhydrin staining. Analytical HPLC was conducted on an
Agilent 1200 HPLC system employing a DAD detector and a ZORBAX ECLIPSE XDBꢀC8 (4.6 mm ×
150 mm, 5 µm) column with the following binary solvent systems: System 1: eluent, methanol/0.1% aq
trifluoroacetic acid, 10% methanol for 3 min, to 100% in 15 min, 100% for 6 min, to 10% in 3min, then
10% for 3min, flow rate 5 mL/min, λ = 210 or 254 nm; System 2: CH3CN/0.1% aq trifluoroacetic acid,
10% CH3CN for 3 min, to 100% in 15 min, 100% for 6 min, to 10% in 3min, then 10% for 3min, flow
rate 5 mL/min, λ = 210 or 254 nm. Preparative HPLC was performed on an Agilent 1100 Preparative
Series, using a MACHEREYꢀNAGEL Varioprep VP 250/32 Nucleodur C18 HTec (32 x 250 mm, 5
ꢂm, flow rate 32 mL/min), a ZORBAX ECLIPSE XDBꢀC8 PrepHT (21.5 x 150 mm, 5 ꢂm, flow rate
10 mL/min) or a MACHEREYꢀNAGEL Varioprep VP 250/10 Nucleodur C18 HTec (10 x 250 mm, 5
ꢂm, flow rate 3 mL/min) column with the solvent systems indicated.
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
2,8ꢀBis(benzyloxy)ꢀ5ꢀbromoquinoline (5d). To a solution of 5c (863 mg, 2.6 mmol) and
K2CO3 (1.08 g, 2.6 mmol) in DMF (26 mL) was added dropwise benzyl bromide (0.5 mL, 3.9 mmol).
The mixture was stirred at room temperature for 2 h. The solvent was evaporated and the residue was
diluted with water. The product was extracted with EtOAc, the combined organic layers were washed
with 0.5 M HCl, water and brine, dried (Na2SO4) and evaporated. Purification by flash chromatography
1
(hexanes/EtOAc 30:1) yielded 5d (990 mg, 91%) as a colorless solid. H NMR (CDCl3, 600 MHz) δ
5.31 (s, 2H), 5.58 (s, 2H), 6.96 (d, J = 8.4 Hz, 1H), 7.05 (d, J = 9.1 Hz, 1H), 7.28 – 7.36 (m, 4H), 7.37
ACS Paragon Plus Environment

Page 25 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
13
– 7.42 (m, 2H), 7.50 (d, J = 8.4 Hz, 1H), 7.52 – 7.58 (m, 4H), 8.33 (d, J = 9.1 Hz, 1H). C NMR
(CDCl3, 151 MHz) δ 68.16, 71.55, 112.9, 113.0, 114.7, 125.4, 127.4, 127.5, 128.0, 128.1, 128.5, 128.7,
128.8, 137.2, 137.2, 138.7, 139.5, 153.2, 161.6. IR (NaCl) ν 3030, 1610, 1597, 1502, 1404, 1272, 1253,
1095, 1006, 914, 826, 732, 695 cmꢀ1. HPLC (system 1) tR = 23.1 min, purity >99 %, (system 2) tR =
22.6 min, purity >99 %. ESIꢀMS m/z 420.0
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
tertꢀButyl 4ꢀ[2,8ꢀbis(benzyloxy)quinolinꢀ5ꢀyl]piperazineꢀ1ꢀcarboxylate (5e). To a solution of
1ꢀbocꢀpiperazine (1.3 g, 7.2 mmol) in toluene (51 mL) were added Pd(OAc)2 (32 mg, 0.14 mmol) and
racꢀBINAP (286 mg, 0.43 mmol) under N2ꢀatmosphere. After stirring for 15 min at room temperature,
5d (1.5 g, 3.6 mmol) and NaOtBu (0.5 g, 5.0 mmol) were added and the resulting mixture was stirred
under reflux for 24 h. After cooling down to room temperature, EtOAc was added and the organic layer
was washed with 0.5 M HCl, water and brine, dried (Na2SO4) and evaporated. Purification by flash
1
chromatography (hexanes/EtOAC 20:1 to 2:1) yielded 5e (1.2 g, 63%) as a colorless solid. H NMR
(CDCl3, 600 MHz) δ 1.50 (s, 9H), 2.79 – 3.08 (m, 4H), 3.37 – 3.77 (m, 4H), 5.30 (s, 2H), 5.58 (s, 2H),
6.88 (d, J = 8.3 Hz, 1H), 6.98 (d, J = 9.0 Hz, 1H), 7.05 (d, J = 8.3 Hz, 1H), 7.28 – 7.33 (m, 2H), 7.33 –
7.40 (m, 4H), 7.47 – 7.67 (m, 4H), 8.39 (d, J = 9.0 Hz, 1H). 13C NMR (CDCl3, 91 MHz) δ 28.62,
53.47, 67.85, 71.98, 79.99, 112.8, 113.0, 113.5, 122.2, 127.4, 127.8, 128.0, 128.5, 128.6, 128.7, 135.2,
137.6, 137.9, 139.7, 143.6, 150.3, 155.0, 161.3. IR (NaCl) ν 2924, 2854, 1695, 1610, 1411, 1261, 1246,
1169, 1003, 734, 696 cmꢀ1. HPLC (system 1) tR = 23.1 min, purity >99 %, (system 2) tR = 22.5 min,
purity >99 %. ESIꢀMS m/z 526.7 [M+ H+].
tertꢀButyl 4ꢀ[2,8ꢀbis(benzyloxy)quinolinꢀ5ꢀyl]ꢀ1,4ꢀdiazepaneꢀ1ꢀcarboxylate (5f). To a
solution of Nꢀ(tertꢀbutylꢀoxycarbonyl)ꢀhomopiperazine (380 mg, 1.9 mmol) in toluene (15 mL),
Pd(OAc)2 (8.5 mg, 0.04 mmol) and racꢀBINAP (71 mg, 0.11 mmol) were added. After stirring for
15 min, 5d (400 mg, 0.95 mmol) and NaOtBu (128 mg, 1.3 mmol) were added. The tube was sealed
and the reaction mixture was stirred at 115°C for 1 d. After cooling down to room temperature, the
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 68
1
2
3
4
5
6
7
8
9
reaction mixture was diluted with EtOAc, the organic layer was washed with 0.5 M HCl, water and
brine, dried (Na2SO4) and evaporated. Purification by flash chromatography (hexanes/EtOAc 40:1 to
10:1 to 2:1) yielded 1f (350 mg, 68%) as a colorless oil. 1H NMR (CDCl3, 600 MHz) δ 1.50 (d, J = 9.8
Hz, 9H), 1.86 – 2.13 (m, 2H), 3.02 – 3.27 (m, 4H), 3.55 – 3.73 (m, 4H), 5.30 (s, 2H), 5.58 (s, 2H), 6.94
(dd, J = 8.2, 2.6 Hz, 1H), 6.98 (d, J = 9.0 Hz, 1H), 7.00 – 7.06 (m, J = 8.2, 2.0 Hz, 1H), 7.28 – 7.33 (m,
2H), 7.33 – 7.41 (m, 4H), 7.55 (dd, J = 11.8, 7.4 Hz, 4H), 8.44 (d, J = 8.8 Hz, 1H). 13C NMR (CDCl3,
151 MHz) δ 28.69, (29.19), 29.35, (45.69), 46.48, (47.79), 48.14, (56.57), 56.97, (57.40), 57.46, 67.80,
71.98, (79.60), 79,64, (112.7), 113.1, (115.2), 115.2, 122.8, 127.4, 127.8, 128.0, 128.5, 128.6, 128.7,
135.6, (137.6), 137.6, 137.9, 139.5, (146.0), 146.1, (150.0), 150.0, (155.7), 155.8, 161.3 (signals in
brackets represent resonances caused by rotamers). IR (NaCl) ν 3432, 2974, 1645, 1481, 1453, 1408,
1366, 1312, 1261, 1154, 736, 699 cmꢀ1. HPLC (system 1) tR = 23.0 min, purity >98 %, (system 2) tR =
22.7 min, purity >99 %. ESIꢀMS m/z 540.2 [M+ H+].
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
4ꢀ[2,8ꢀBis(benzyloxy)quinolinꢀ5ꢀyl]piperazine (5g). To a solution of 5e (550 mg, 0.95 mmol)
in CH2Cl2 (13 mL) was added TFA (1.3 mL). The mixture was stirred at room temperature for 2 h. The
reaction was stopped by adjusting the pH to 9 by addition of 1 M NaOH. The product was extracted
with CH2Cl2 and the combined organic phases were dried over Na2SO4 and evaporated. Purification by
1
flash chromatography (CH2Cl2/MeOH 15:1) yielded 5g (220 mg, 51 %) as a yellow solid. H NMR
(CDCl3, 360 MHz) δ 3.11 (m, 4H), 3.26 (m, 4H), 3.45 (s, 1H), 5.31 (s, 2H), 5.58 (s, 2H), 6.94 (d, J =
8.3 Hz, 1H), 6.99 (d, J = 9.0 Hz, 1H), 7.06 (d, J = 8.4 Hz, 1H), 7.43 – 7.27 (m, 6H), 7.55 (t, J = 7.3 Hz,
13
4H), 8.36 (d, J = 9.1 Hz, 1H). C NMR (CDCl3, 91 MHz) δ 45.54, 52.64, 67.89, 71.89, 112.8, 113.0,
114.0, 122.1, 127.4, 127.9, 128.0, 128.5, 128.6, 128.7, 135.0, 137.5, 137.8, 139.7, 143.0, 150.6, 161.3.
IR (NaCl) ν 3439, 3031, 2946, 1610, 1411, 1264, 1005, 742, 693 cmꢀ1. HPLC (system 1) tR = 19.9 min,
purity 99 %, (system 2) tR = 16.5 min, purity 93 %. ESIꢀMS m/z 426.1 [M+ H+].
ACS Paragon Plus Environment

Page 27 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
4ꢀ[2,8ꢀBis(benzyloxy)quinolinꢀ5ꢀyl]ꢀ1,4ꢀdiazepane (5h). Compound 5h was prepared as
described for 5g, using a mixture of 5f (200 mg, 0.37 mmol) and TFA (0.5 mL) in CH2Cl2 (5.0 mL).
Purification by flash chromatography (CH2Cl2/MeOH 15:1) yielded 5h (129 mg, 79%) as a yellow
7
8
9
1
solid. H NMR (CDCl3, 600 MHz) δ 1.95 – 2.12 (m, 2H), 3.09 – 3.40 (m, 8H), 5.30 (s, 2H), 5.58 (s,
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
2H), 6.94 – 7.01 (m, 2H), 7.04 (d, J = 8.4 Hz, 1H), 7.27 – 7.32 (m, 2H), 7.33 – 7.40 (m, 4H), 7.49 –
7.58 (m, 4H), 8.46 (d, J = 9.0 Hz, 1H). 13C NMR (CDCl3, 91 MHz) δ 29.97, 46.92, 49.40, 56.66, 57.79,
67.82, 72.02, 112.8, 113.1, 115.4, 122.8, 127.4, 127.8, 127.9, 128.5, 128.6, 128.7, 135.6, 137.6, 137.9,
139.6, 145.9, 150.1, 161.3. IR (NaCl) ν 3432, 3034, 2937, 2848, 1650, 1559, 1506, 1411, 1260, 1093,
733, 697 cmꢀ1. HPLC (system 1) tR = 20.0 min, purity 87 %, (system 2) tR = 16.8 min, purity 91 %.
ESIꢀMS m/z 440.2 [M+ H+].
tertꢀButyl
4ꢀ[8ꢀ(benzyloxy)ꢀ2ꢀoxoꢀ2Hꢀchromenꢀ5ꢀyl]piperazineꢀ1ꢀcarboxylate
(6d).
Pd2(dba)3 (24 mg, 0.03 mmol) and racꢀBINAP (49 mg, 0.08 mmol) were dissolved in toluene (9 mL)
and stirred for 5 min under nitrogen atmosphere. Then Cs2CO3 (300 mg, 0.92 mmol), 1ꢀbocꢀpiperazine
(307 mg, 1.6 mmol), and 6c (220 mg, 0.66 mmol) were added and the reaction mixture was stirred in a
sealed microwave reaction tube at 85°C for 1 d. After cooling down to room temperature, the reaction
was diluted with EtOAc was added. The combined organic layer was washed with 0.5 M HCl, water
and brine, dried (Na2SO4) and evaporated. Purification by flash chromatography (hexanes/EtOAc 20:1
to 5:1) yielded 6d (172 mg, 59%) as a yellow solid. 1H NMR (CDCl3, 600 MHz) δ 1.49 (s, 9H), 2.75 –
3.08 (m, 4H), 3.45 – 3.88 (m, 4H), 5.20 (s, 2H), 6.43 (d, J = 9.8 Hz, 1H), 6.80 (d, J = 8.7 Hz, 1H), 7.03
(d, J = 8.7 Hz, 1H), 7.32 (t, J = 7.3 Hz, 1H), 7.38 (t, J = 7.5 Hz, 2H), 7.47 (d, J = 7.4 Hz, 2H), 8.07 (d,
13
J = 9.8 Hz, 1H). C NMR (CDCl3, 91 MHz) δ 28.58, 44.16, 53.55, 71.85, 80.20, 114.7, 115.4, 116.0,
116.9, 127.5, 128.2, 128.8, 136.6, 140.1, 143.0, 143.9, 145.5, 154.9, 160.3. IR (NaCl) ν 2975, 2917,
2857, 1727, 1690, 1487, 1454, 1424, 1267, 1250, 1170, 1127, 1069, 1011, 908, 810 cmꢀ1. HPLC
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 68
1
2
3
(system 1) tR = 21.2 min, purity 94 %; (system 2) tR = 19.8 min, purity 99 %. ESIꢀMS m/z 459.2
4
5
[M+Na+].
6
7
8
tertꢀButyl 4ꢀ[8ꢀ(benzyloxy)ꢀ2ꢀoxoꢀ2Hꢀchromenꢀ5ꢀyl]ꢀ1,4ꢀdiazepaneꢀ1ꢀcarboxylate
(6e).
9
Compound 6e was prepared according to the protocol of 6d, using a solution of Pd2(dba)3 (24 mg,
0.03 mmol), racꢀBINAP (49 mg, 0.08 mmol), Cs2CO3 (300 mg, 0.92 mmol), Nꢀ(tertꢀbutylꢀ
oxycarbonyl)ꢀhomopiperazine (307 mg, 1.6 mmol), and 6c (220 mg, 0.66 mmol) in toluene (9 mL).
Purification by flash chromatography (hexanes/EtOAc 20:1 to 5:1) yielded 6e (172 mg, 59%) as a
yellow solid. 1H NMR (CDCl3, 600 MHz) δ 1.46 – 1.53 (m, 9H), 1.88 – 2.04 (m, 2H), 3.03 – 3.16 (m,
4H), 3.52 – 3.59 (m, 2H), 3.59 – 3.70 (m, 2H), 5.20 (s, 2H), 6.41 (d, J = 9.7 Hz, 1H), 6.85 (dd, J = 8.7,
4.0 Hz, 1H), 7.00 (dd, J = 8.6, 3.6 Hz, 1H), 7.31 – 7.33 (m, 1H), 7.36 – 7.40 (m, 2H), 7.45 – 7.48 (m,
2H), 8.07 – 8.14 (m, 1H). 13C NMR (CDCl3, 101 MHz) δ 28.50, (28.85), 29.08, (45.46), 46.33,
(47.25), 47.66, (56.34), 56.75, 57.19, 71.67, 115.6, 116.3, 116.8, 127.4, 128.0, 128.6, 129.5, 134.0,
136.5, 140.5 (signals in brackets represent resonances caused by rotamers) IR (NaCl) ν 3416, 2974,
2927, 1730, 1655, 1487, 1264, 1070 cmꢀ1. HPLC (system 1) tR = 21.5 min, purity 96 %; (system 2) tR =
20.1 min, purity 96 %. ESIꢀMS m/z 451.1 [M+H+].
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
4ꢀ[8ꢀ(Benzyloxy)ꢀ2ꢀoxoꢀ2Hꢀchromenꢀ5ꢀyl]piperazine (6f). Compound 6f was prepared as
described for 6g, using a mixture of 6d (170 mg, 0.39 mmol) and TFA (0.3 mL) in CH2Cl2 (5.3 mL).
Purification by flash chromatography (CH2Cl2/MeOH 15:1) yielded 6f (85 mg, 65%) as a yellow solid.
1H NMR (DMSOꢀd6, 360 MHz) δ 2.71 – 2.87 (m, 4H), 2.87 – 3.03 (m, 4H), 5.20 (s, 2H), 6.46 (d, J =
9.8 Hz, 1H), 6.93 (d, J = 8.8 Hz, 1H), 7.30 (d, J = 8.8 Hz, 1H), 7.34 – 7.38 (m, 1H), 7.38 – 7.45 (m,
13
2H), 7.46 – 7.52 (m, 2H), 8.11 (d, J = 9.8 Hz, 1H). C NMR (DMSOꢀd6, 151 MHz) δ 45.53, 54.00,
70.45, 114.2, 114.3, 115.1, 116.3, 127.8, 128.0, 128.5, 136.7, 140.7, 141.6, 144.2, 159.5. IR (NaCl) ν
3432, 1684, 1377, 1272, 1204, 1131, 1060 cmꢀ1. HPLC (system 1) tR = 16.9 min, purity 99 %; (system
2) tR = 14.0 min, purity 99 %. ESIꢀMS m/z 337.2 [M+H+].
ACS Paragon Plus Environment

Page 29 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
4ꢀ[8ꢀ(Benzyloxy)ꢀ2ꢀoxoꢀ2Hꢀchromenꢀ5ꢀyl]ꢀ1,4ꢀdiazepane (6g). Compound 6g was prepared
as described for 5g, using a mixture of 6e (87 mg, 0.19 mmol) and TFA (0.25 mL) in CH2Cl2
(2.50 mL). Purification by flash chromatography (CH2Cl2/MeOH 15:1) yielded 6g (38 mg, 57%) as a
yellow solid. 1H NMR (CDCl3, 400 MHz) δ 1.91 – 2.02 (m, 2H). 3.08 – 3.13 (m, 2H), 3.13 – 3.21 (m,
2H), 3.22 – 3.28 (m, 2H), 3.29 – 3.41 (m, 2H), 5.19 (d, J = 8.3 Hz, 1H), 6.48 (d, J = 9.8 Hz, 1H), 7.04
(d, J = 8.8 Hz, 1H), 7.31 (d, J = 8.9 Hz, 1H), 7.34 – 7.38 (m, 1H), 7.39 – 7.45 (m, 2H), 7.46 – 7.52 (m,
2H), 8.24 (d, J = 9.8 Hz, 1H). 13C NMR (CDCl3, 101 MHz) δ 27.33, 45.12, 47.02, 54.48, 55.86, 70.42,
115.0, 115.2, 116.3, 116.5, 127.8, 128.0, 128.5, 136.6, 141.2, 141.7, 143.9, 145.3, 159.5. IR (NaCl) ν
3421, 2959, 2922, 2850, 1722, 1651, 1488, 1454, 1270, 1192, 1134, 1067, 806, 699 cmꢀ1. HPLC
(system 1) tR = 17.1 min, purity 91 %; (system 2) tR = 14.2 min, purity 88 %. ESIꢀMS m/z 351.1
[M+H+].
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
4ꢀBenzylꢀ5ꢀ(benzyloxy)ꢀ8ꢀbromoꢀ2Hꢀbenzo[b][1,4]oxazinꢀ3(4H)ꢀone (7d). To a solution of
7c (3.7 g, 11.0 mmol) and K2CO3 (4.5 g, 33.0 mmol) in DMF (110 mL) was added dropwise benzyl
bromide (1.9 mL, 16.5 mmol). The mixture was stirred at room temperature for 2 h. The solvent was
evaporated and the residue was diluted with water. The product was extracted with EtOAc, the
combined organic layers were washed with 0.5 M HCl, water and brine, dried (Na2SO4) and
evaporated. Purification by flash chromatography (hexanes/EtOAc 12:1 to 1:1) yielded 7d (2.67 g,
1
57%) as a colorless solid. H NMR (DMSOꢀd6, 600 MHz) δ 4.65 (s, 2H), 4.95 (s, 2H), 5.37 (s, 2H),
6.50 (d, J = 9.0 Hz, 1H), 6.99 (d, J = 7.1 Hz, 2H), 7.13 (d, J = 9.0 Hz, 1H), 7.14 – 7.20 (m, 3H), 7.28 –
7.32 (m, 2H), 7.34 – 7.41 (m, 3H). 13C NMR (DMSOꢀd6, 151 MHz) δ 46.89, 69.24, 71.63, 102.8,
109.2, 120.2, 127.0, 127.1, 127.7, 128.2, 128.4, 128.6, 128.9, 135.8, 137.0, 146.7, 148.8, 166.2. IR
(NaCl) ν 1692, 1485, 1456, 1386, 1279, 1103, 1016, 739, 702 cmꢀ1. HPLC (system 1) tR = 21.7 min,
purity 89 % (system 2) tR = 20.2 min, purity 87 %. ESIꢀMS m/z 425.9 [M+H+].
ACS Paragon Plus Environment