Hybridization of β-Adrenergic Agonists and Antagonists Confers G Protein Bias

Article abstract of DOI:10.1021/acs.jmedchem.9b00349

Starting from the β-adrenoceptor agonist isoprenaline and beta-blocker carvedilol, we designed and synthesized three different chemotypes of agonist/antagonist hybrids. Investigations of ligand-mediated receptor activation using bioluminescence resonance energy transfer biosensors revealed a predominant effect of the aromatic head group on the intrinsic activity of our ligands, as ligands with a carvedilol head group were devoid of agonistic activity. Ligands composed of a catechol head group and an antagonist-like oxypropylene spacer possess significant intrinsic activity for the activation of Gα_s, while they only show weak or even no β-arrestin-2 recruitment at both β₁- and β₂-AR. Molecular dynamics simulations suggest that the difference in G protein efficacy and β-arrestin recruitment of the hybrid (S)-22, the full agonist epinephrine, and the β₂-selective, G protein-biased partial agonist salmeterol depends on specific hydrogen bonding between Ser^5.46 and Asn^6.55, and the aromatic head group of the ligands.

Full text of DOI:10.1021/acs.jmedchem.9b00349

Subscriber access provided by UNIV AUTONOMA DE COAHUILA UADEC
Article
Hybridization of #-adrenergic agonists and antagonists confers G protein bias
Markus Stanek, Louis-Philippe picard, Maximilian F. Schmidt, Jonas M.
Kaindl, Harald Hübner, Michel Bouvier, Dorothee Weikert, and Peter Gmeiner
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.9b00349 • Publication Date (Web): 01 May 2019
Downloaded from http://pubs.acs.org on May 2, 2019
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Hybridization of β-adrenergic agonists and antagonists
confers G protein bias
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
1
1
1
Markus Stanek, Louis-Philippe Picard, Maximilian F. Schmidt, Jonas M. Kaindl, Harald Hübner,
2*
1*
1*
Michel Bouvier, Dorothée Weikert, Peter Gmeiner
1Department of Chemistry and Pharmacy, Medicinal Chemistry, Friedrich-Alexander University Erlangen-
Nürnberg (FAU), Nikolaus-Fiebiger-Str. 10, 91058 Erlangen, Germany
2Institute for Research in Immunology and Cancer (IRIC), Department of Biochemistry and Molecular Medicine,
University of Montreal, Québec, Canada H3C 3J7
ABSTRACT
Starting from the β-adrenoceptor agonist isoprenaline and the beta-blocker carvedilol, we
designed and synthesized three different chemotypes of agonist/antagonist hybrids. Investigations of
ligand-mediated receptor activation using BRET-biosensors revealed a predominant effect of the
aromatic head group on the intrinsic activity of our ligands, as ligands with a carvedilol head group were
devoid of agonistic activity. Ligands composed of a catechol head group and an antagonist-like
oxypropylene-spacer possess significant intrinsic activity for the activation of Gα , while they only show
s
weak or even no β-arrestin-2 recruitment at both, β - and β -AR. Molecular dynamics simulations suggest
1
2
that the difference in G protein efficacy and β-arrestin recruitment of the hybrid (S)-22, the full agonist
epinephrine and the β -selective, G protein-biased partial agonist salmeterol depends on specific
2
hydrogen bonding between Ser5.46, Asn6.55 and the aromatic head group of the ligands.
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
INTRODUCTION
Beta adrenergic receptors (β-AR) undoubtedly belong to the most thoroughly studied members
within the superfamily of G protein coupled receptors (GPCRs). They are involved in the regulation of
important physiological functions and it is therefore not surprising that numerous drugs targeting these
receptors have been developed. β-AR antagonists (β-blockers) are used in the treatment of heart failure
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
and angina pectoris, where they reduce myocardial demand and workload through inhibition of the
1
endogenous catecholamines at the β -AR subtype. In contrast, β-AR agonists reduce smooth muscle
1
contraction in the airways by interaction with the β -AR subtype and can therefore improve the symptoms
2
of obstructive respiratory diseases such as asthma and chronic obstructive pulmonary disease (COPD).2,
3
Structurally, β-adrenoceptor agonists and antagonists share a number of common features. Both are
composed of an aromatic head group that is connected to a second, very often lipophilic moiety (“tail
region”) through an aminoalcohol linker. The linker consists of an aminoethanol moiety for adrenoceptor
agonists, but formal introduction of an oxymethylene unit results in a connection by oxy-1,2-
propanolamine-linkers, typical for the class of β-adrenoceptor blockers.4
2
Figure 1. The comparison of the carazolol-, adrenaline- and salmeterol-bound crystal structures of β -AR reveals highly
similar interactions of the aminoalcohol chain and polar residues in TMs 3 and 7 but major differences in the hydrogen bond
network between the ligands and TMs 5 and 6.
Structural information on the ligand binding mode for agonists and antagonists is available
because high resolution crystal structures of both, β -AR and β -AR have been determined. A thermo-
1
2
2
ACS Paragon Plus Environment

Page 3 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
stabilized form of the turkey β -AR has been crystallized in the presence of a number of different ligands
1
including agonists, partial agonists and antagonists.5-7 Crystal structures of β -AR in the inactive form
2
were also obtained in the presence of agonists and antagonists,8-11 but active state structures became
available when heterotrimeric complexes of receptor and agonist together with the G protein or a suitable
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
G protein mimic were constructed.12-15 Very recently, a crystal structure of β -AR in complex with a
2
camelid antibody (nanobody, Nb71) and the G protein-biased partial agonist salmeterol was also
16
released. These crystal structures reveal a highly similar arrangement within the orthosteric binding
pocket of the receptor ligand complexes (Figure 1). All ligands form hydrogen bonds between the
aminoalcohol linker and residues Asp3.32, Asn7.39 and Tyr7.43 (superscript indicates Ballesteros-Weinstein
1
7
numbering ) located at the front part of the binding site formed by TM3 and TM7. These interactions
anchor the ligand within the receptor and are conserved among most β-adrenergic ligands.6, 7, 11 The
second part of the orthosteric binding pocket is comprised by Ser5.42, Ser5.46 and Asn6.55. Specific
interactions of the polar catechol (mimetic) unit of β-adrenoceptor agonists with these hydrophilic
residues are able to induce a small contraction (1-3 Å) of the binding cavity, most notably by an inward
movement and rotamer change of Ser5.46. This contraction is followed by the reorganization of the core
triad residues (Pro5.50, Ile3.40 and Phe6.44) which, together with the allosteric effect of the intracellular
binding partner (G protein or nanobody), stabilizes the active receptor state including the large outward
movement of the cytoplasmic end of TM6.6, 12-14 β-adrenoceptor antagonists bind slightly different to this
region, because the greater spatial requirement of the aromatic system and the lack of polar groups
together with the increased molecular distance between the aminoalcohol function and the aromatic head
group (~2 Å) prevent sufficient stabilization of the contracted binding pocket.6, 8, 11
The importance of an intact hydrogen bond network for receptor activation has been demonstrated
by mutational studies.16, 18, 19 Furthermore, recent crystal structures and molecular dynamics simulations
of biased and/or partial agonists have implicated alterations in these hydrogen bond networks in the origin
5
, 6, 16, 20
of submaximal intrinsic activity and alternate signaling signatures at β - and β -AR.
Alternate
3
1
2
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
signaling signatures or so called biased signaling has gained increasing interest over the past decade. It
is thought that by preferential activation of one signaling pathway over the other, safer therapies may be
developed, if distinct pathways governed by the same receptor are involved in desired versus adverse
21
effects. Although the physiological consequences of functionally selective activation of β-AR are far
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
from being fully understood, evidence favoring the activation of certain signaling pathways over others,
depending on the involved tissue, has accumulated over recent years.21-23 For instance, carvedilol, an
24
inverse agonist of Gα -activation and a weak agonist for β-arrestin-dependent ERK phosphorylation,
s
has been shown to possess mortality advantage over other β-blockers in the treatment of heart failure. It
has been suggested that this effect may be due to a cardioprotective effect of β -AR and β -AR mediated
1
2
2
4
arrestin-dependent signaling in the heart. On the contrary, prolonged agonist stimulation of β -AR in
2
the airways is associated with receptor desensitization leading to a reduced responsiveness to short acting
β2-AR agonists which are used as rescue medication for asthma. Moreover, stimulation of β-arrestin
2
5
recruitment in airway epithelial cells has been implicated as proinflammatory in asthma models. Thus,
the use of Gα -biased β -AR ligands has been suggested to be beneficial for the treatment of obstructive
s
2
pulmonary diseases.26
Taking advantage of the recent findings from crystal structures on the tight but yet different
hydrogen bond networks of β-AR agonists and antagonists (Figure 1), we aimed for the design, synthesis
and pharmacological characterization of novel β-adrenoceptor ligands based on a combination of agonist
and antagonist structural motifs. Ideally such compounds should address only a subset of the
agonist/receptor contacts, thereby making them ideal candidates for developing ligands with alternate
signaling profiles.
RESULTS AND DISCUSSION
Design. In an effort directed to an investigation of individual contributions of the spacer region,
the hydrophobic tail substituent and the orthosteric head group to the recognition and possibly biased
4
ACS Paragon Plus Environment

Page 5 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
activation of β-adrenergic receptors, we designed type I-III hybrid ligands starting from the reference
agonist isoprenaline, the β-blocker carvedilol and our recently developed high affinity β -agonist
2
1
0
FAUC72 (Figure 2). These ligands do not only differ in their aromatic head group and length of the
spacer region but also in the presence of additional lipophilic amino substituents. Very recently, we could
demonstrate that addition of such lipophilic moieties to the primary or secondary amine of the
endogenous catecholamines adrenaline, noradrenaline and dopamine was a successful means of
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
27
generating high affinity dual adrenergic/dopaminergic receptor agonists ((R)-1, (R)-2, Table 1). Thus,
hybrid compounds of type I were composed of the 9H-carbazolyl head group and the 1-ethoxy-2-
methoxy-phenyl tail of carvedilol, but they also shared the typical ethanolamine substructure of the β-
1
0
adrenergic agonists isoprenaline and FAUC72 .
O
OH
H
HN
N
O
type I hybrid
O
OH
OH
O
O
H
H
OH
HO
HO
N
HN
HO
N
O
O
H
HN
O
N
(
S)-carvedilol
(
R)-isoprenaline
FAUC72
OH
O
H
HO
HO
N
R
O
HO
HO
O
O
R =
OH
H
type II hybrid
O
O
O
N
type III hybrid
Figure 2. Design of agonist-antagonist hybrid adrenoceptor ligands starting from the lead structures (R)-isoprenaline, (S)-
carvedilol and FAUC72. Type I-III hybrids are designed by combining typical agonist and antagonist moieties: catechol or
catechol mimetic head groups are shown in blue, the prototypical agonist ethanolamine-linker in orange and the lipophilic
amino-substituents in red. Typical antagonist scaffolds differ in the head group (carbazol, dark green) and possess an
oxymethyl-insertion in the linker region (light green).
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
In contrast, hybrid compounds of type II bear the classical catechol head group, but are connected
to the individual tail substituents via the antagonist-like oxypropanolamine linker. Similar hydroxylated
phenoxypropanolamine derivatives with isopropyl- or tertiary butyl- substituents have already been
described as potential β-adrenergic agonists in the late 1970s. It was found that these ligands are able to
stimulate β-adrenergic receptors in atrial and tracheal tests, if they were substituted with hydroxyl groups
in the correct positions.28, 29 However, a dissection of the involved signaling pathways and individual
contributions of the molecule substructures is still lacking. In addition to the hybrid agonist/antagonist
ligands of type II, we also aimed at understanding the overall influence of the spacer length between the
aromatic core moiety and the basic amine. Therefore, a formal methylene-introduction was performed
yielding type III hybrid as homologous catecholamine comprising a 1,2-propanolamine substructure. All
hybrid ligands were synthesized with their central hydroxyl in respective (R)- and (S)- configuration and
their deshydroxy-form devoid of that hydroxyl to enable a variable degree of engagement in hydrogen
bonding to the receptors.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Synthesis. Compared to the β-blocker carvedilol, type I hybrids differ in the length and nature of
their linker. Most importantly, the absence of the phenolic ether function requires the synthesis to proceed
via a carbon-carbon connecting reaction in order to attach the side chain to the aromatic head group. To
this end, we chose a Stille cross coupling reaction, which we had successfully applied earlier for the
2
7
synthesis of β -AR agonists with benzoxazinone headgroups. Similar to a literature protocol for the
2
synthesis of (R)-/(S)-carvedilol and its deshydroxy analogue 3, we could start from 4-hydroxy-9H-
carbazole (4), which was activated by reaction with triflic anhydride. The unexpectedly stable triflate 5
was then used for the Stille-coupling with tributyl-cis-(ethoxyethenyl)stannane affording enolether 6.
Subsequent acidic hydrolysis yielded the carbazolylacetaldehyde 7 that was used for the next reaction
step without further purification. The required building block 2-(2-methoxyphenoxy)ethylamine (8) was
prepared starting from 2-methyloxazoline that was heated up together with 2-methoxyphenol. The
6
ACS Paragon Plus Environment

Page 7 of 71
Journal of Medicinal Chemistry
1
2
3
resulting N-acetylated intermediate 9 was purified by recrystallization and subsequently deacylated in
3
0
5
M HCl yielding the primary amine 8 in high purity after aqueous workup (Scheme 1). Reductive
amination of 7 with 8 in presence of sodium triacetoxy borohydride resulted in the final
carbazolylethylamine 10.
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The synthesis of the two carbazolylethanolamines (R)-15 and (S)-15 proceeded in a very similar
way. When tributyl-ethenylstannane instead of tributyl-cis-(ethoxyethenyl)stannane was used for the
Stille reaction with intermediate 5, 4-vinyl-9H-carbazole (11) was obtained in excellent yield of 87%.
3
1
Sharpless asymmetric dihydroxylation using AD mix-α or –β yielded the chiral diols (S)/(R)-12, which
were converted into the mono-tosylates (S)/(R)-13 by regioselective O-tosylation at the primary hydroxyl
3
2
using p-toluenesulfonylchloride in the presence of dibutyltinoxide. Treatment of these intermediates
with sodium hydroxide in DMF gave access to the chiral epoxides (S)/(R)-14. The ethanolamines (S)/(R)-
1
5 were obtained in excellent enantiopurity (ee 95%) in a final epoxide opening with the primary amine
8. For comparison, we also synthesized the individual isomers of carvedilol and its deshydroxy analogue
33
3
. Alkylation of 4 with (+)-or (-)-epichlorohydrin or 1,3-dibromopropane yielded intermediates 44 or
(S)-/(R)-45, respectively, which were subsequently reacted with the nucleophile 8 to give (R)-/(S)-
carvedilol (ee 95-96 %) and 3 in good yield.
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Scheme 1. Synthesis of 9H-carbazol-derived ligands.a
O
O
O
N
R2
a or b
c or d
Br
H
HN
HN
OH
O
HN
O
N
m
R1
O
44: R =
(S)-carvedilol: R = (S)-OH
2
R)-carvedilol: R = (R)-OH
H
4
1
N
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
O
(
R4
O
e
2
(
S)-45 / (R)-45: R =
*
1
3
: R = H
9: R = acetyl
4
2
n
8: R = H
4
O
O
f
g
h
H
HN
HN
HN
O
HN
N
OTf
O
5
6
7
10
i
O
O
j
OH

l
O
OH

d
H
N

HN
HN
HN
HN
OR3
11
(S)-12 / (R)-12 : R3 = H
(S)-13 / (R)-13 : R = Tos
(S)-14 / (R)-14
(S)-15 / (R)-15
k
3
a
Reagents and conditions: a) 1,3-dibromopropane, acetone, K
acetone, K CO , reflux, 48 h, 57-58%; c) 8, KI, CH CN, N , reflux, 5 h, 73%; d) 8, isopropanol, reflux, 4 – 24 h, 34-70%; e)
trifluoromethanesulfonic anhydride, triethylamine, CH Cl , -10°C to rt, 1 h, 84%; f) (PPh Pd(II)Cl , LiCl, DMF,
Bu (ethoxyethenyl)Sn, Ar, 90°C, 3 h, 90%; g) 2 M HCl, reflux, 0.5 h, crude; h) 8, Na(OAc) BH, THF, rt, Ar, 2.5 h, 36% (two
steps); i) (PPh Pd(II)Cl , LiCl, DMF, Bu (ethenyl)Sn, Ar, 90°C, 2 h, 87%; j) AD mix-β or –α, tert.-butanol/H O 1:1, 0°C,
h, 79-84%; k) p-toluenesulfonyl chloride, SnOBu , triethylamine, CH Cl /CH CN 2:1, Ar, 0°C, 4 h, 70-86%; l) NaOH,
2 3
CO , reflux, 16 h, 76%; b) (R)-(-)- or (S)-(+)-epichlorohydrin,
2
3
3
2
2
2
3
)
2
2
3
3
3
)
2
2
3
2
4
2
2
2
3
DMF, rt, 3 h, 82-88%; m) 2-methoxyphenol, neat, Ar, 160°C, 40 h, 57%; n) 5 M HCl, reflux, 24 h, 100%.
Since ligands with a catechoxypropanolamine substructure have already been described,34, 35 we
35
could adapt a previously established protocol towards the synthesis of our type II agonist/antagonist
hybrids (R)- and (S)-22, -26 and -30 and their respective deshydroxy-derivatives 24, 28 and 32 (Scheme
2). First, 3,4-dihydroxybenzaldehyde (16) was reacted with benzylbromide to prevent the catechol from
oxidation. The resulting benzyl-protected aldehyde 17 was then converted into the phenol 18 by a Baeyer-
Villiger reaction with meta-chloroperbenzoic acid (mCPBA) and subsequent hydrolysis under basic
8
ACS Paragon Plus Environment

Page 9 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
35
conditions. Reaction with enantiomerically pure (R)- or (S)-epichlorohydrin resulted in the chiral
epoxides (S)- and (R)-19. It is worth noting that this reaction proceeds under inversion of the absolute
configuration, as the nucleophilic attack occurs at the epoxide and chloride is released only in the second
36
step. Subsequent epoxide ring-opening reactions were performed using three different primary amines:
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
isopropylamine, 2-(2-methoxyphenoxy)ethylamine (8) or 2-(3-methoxy-4-(3-phenylpropoxy)phenyl)
1
0
ethylamine (33), representing the amino-substituents of known β-adrenoceptor ligands isoprenaline,
10
10
carvedilol and FAUC72 . Instead of the previously described procedure using vanillin, primary amine
3 was prepared starting from commercially available 3-methoxytyramine which was first reacted with
3
N-(benzyloxycarbonyl)succinimide. After O-alkylation with 3-bromo-1-phenylpropane, hydrogenolytic
cleavage of the Cbz protective group afforded 33 in excellent yield (96 %). After reaction with the
primary amines, the resulting benzyl-protected intermediates ((R)- and (S)-21, -25 and -29) displayed
high stereochemical purity (ee 94-98%) which was determined by chiral HPLC. In the last step, these
intermediates were converted into the target oxypropanolamines ((R) and (S)-22, -26 and -30) by catalytic
hydrogenation. In order to prevent the catechol and secondary amines from oxidation and degradation
e.g. by adrenochrome-type reactions,37, 38 hydrochloric acid was added upon completion of the
debenzylation step, and the product was isolated by lyophilization after filtration of the catalyst. The
respective oxypropylamine analogs were obtained in an analogous manner when 18 was reacted with
1
,3-dibromopropane to the bromopropyl-phenylether 20. Nucleophilic displacement of the bromine with
isopropylamine, 8 or 33, respectively, resulted in the intermediates 23, 27 and 31. In the final step, benzyl
protective groups were removed by reductive cleavage to yield the oxypropylamines 24, 28 and 32.
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Scheme 2. Synthesis of catecholamines with an oxypropylene spacer. a
O
O

O
O
O
O
OH
O
O
O
b,c
d
R1
R1
H
R1
R1
R1
R1
16 : R = H
18 : R = Bn
(S)-19 / (R)-19 : R = Bn
1
1
1
a
1
7 : R = Bn
1
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
f
e
OH

O
O
O
NHR2
O
O
O
Br
R1
R1
R1
R1
H N
2
O
20 : R = Bn
(S)-21 / (R)-21 : R = Bn, R =
1 2
(S)-22 / (R)-22 : R = H, R =
1
g
g
g
f
OH
1
2
h
O
O
(
S)-25 / (R)-25 : R = Bn, R =
1 2
O
O
NHR2
R1
(S)-26 / (R)-26 : R = H, R =
H
N
1 2
O
R1
R3
O
(S)-29 / (R)-29 : R = Bn, R =
1 2
(S)-30 / (R)-30 : R = H, R =
O
O
1
2
O
g
23 : R = Bn, R =
1 2
24 : R = H, R =
1
2
O
O
4
6 : R = Cbz
3
27 : R = Bn, R =
1 2
i,j
g
g
33 : R = H
3
28 : R = H, R =
1 2
3
1 : R = Bn, R =
O
O
1
2
32 : R = H, R =
1
2
a
Reagents and conditions: a) benzylbromide, K
KOH, MeOH, 30 min, 87% (two steps); d) (R)-(-)- or (S)-(+)-epichlorohydrin, ethanol, 1.8 M KOH, rt, 24 h, 63-66%; e) 1,3-
dibromopropane, K CO , reflux, 16 h, 93%; f) isopropylamine, 8 or 33, ethanol or CH CN, N , reflux, 4-5 h, 48-85%; g) H
Pd/C, MeOH or ethanol, rt or 60°C, 2-3.5 h, 50-98%; h) N-(benzyloxycarbonyl)succinimide, NaHCO , THF/water, Ar, rt, 4.5
h; i) 3-bromo-1-phenylpropane, K CO , acetone, reflux, 18h, 67% (two steps); j) H , Pd/C, methanol, rt, 3 h, 96%.
2 3 2 2
CO , DMF, Ar, rt, 2.5-3 h 76-96%; b) mCPBA, CH Cl , 0 °C, 1 h; c) 5 M
2
3
3
2
2
,
3
2
3
2
The type III homologous catecholamines (S)- and (R)-39 and their deshydroxy-derivative 43 were
prepared starting from pyrocatechol (Scheme 3), which was subjected to a Williamson ether synthesis
with allylbromide. In accordance to literature procedures, a subsequent Claisen rearrangement of the
3
9
allylether 34 lead to a mixture of 3- and 4-allylcatechol with yields of 40% and 27%, respectively. After
chromatographic separation of the two isomers, the desired 4-allylcatechol (35) was reacted with
benzylbromide to protect the labile catechol substructure. In analogy to the aforementioned syntheses,
1
0
ACS Paragon Plus Environment

Page 11 of 71
1
Journal of Medicinal Chemistry
we aimed for the preparation of enantiomerically pure target compounds. Unfortunately, Sharpless
asymmetric dihydroxylation of the benzylated intermediate 36, subsequent tosylation of the primary
alcohol and aminolysis with 8 did not result in sufficient enantiomeric purity of (S)-38 (ee 24 %) and this
chiral route was therefore not further pursued. Hence, epoxide 37 was synthesized as a racemic mixture
by Prileschajew reaction of 36 with mCPBA. Aminolysis with the primary amine 8 gave the racemic
secondary amine (RS)-38 that was separated into its enantiomers (R)- and (S)-38 by chiral preparative
HPLC. The absolute configuration was assigned by comparison of the chiral HPLC chromatograms with
those of enantiomerically enriched (S)-38 obtained by the chiral route. The target compounds (R)- and
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(S)-39 were obtained in a final debenzylation step in more than ee 96% stereochemical purity and in
excellent yield of 95-96%. The deshydroxy analogue 43 was prepared by hydroboration of the common
intermediate 36. Workup under basic conditions afforded the primary alcohol 40. After activation with
methanesulfonylchloride, intermediate 41 was subjected to a nucleophilic displacement reaction with 8,
resulting in the benzyl-protected homologous dopamine analog 42. In the final step, deprotection yielded
the free catechol 43 as target compound with a propylamine linker structure.
1
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Scheme 3. Synthesis of adrenoceptor ligands with a homologous catecholamine substructure. a

O
R1O
R O
3
R O
1
O
O
b
d
e
N
H
O
OH
HO
R1O
c
R O
3
R O
1
34
35 : R = H
37 : R = Bn
(S)-38 / (R)-38 : R = Bn
1
1
1
f
(
S)-39 / (R)-39 : R = H
3
6 : R = Bn
1
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
a
R O
R O
O
1
1
e
g
OR2
N
H
HO
HO
R O
1
R O
1
O
4
4
0 : R = Bn, R = H
1 : R = Bn, R = Ms
42 : R = Bn
h
1
2
f
1
43 : R = H
1
2
1
BnO
k,e
i
OR3
(S)-38
(ee 24%)
OH
S)-47 : R = H
BnO
j
(
(
3
S)-48 : R = Tos
3
a
Reagents and conditions: a) allylbromide, K
7% (4-allylcatechol); c) benzylbromide, K CO
, Pd/C, MeOH or ethanol, rt or 60°C, 2-3.5 h, 80-96%; g) 1 M BH
Ar, 0°C to rt, 3 h, followed by 5 M NaOH, 30% H , 0.5 h, 76%; h) methanesulfonyl chloride, triethylamine, CH
h, 92%, i) AD-mix α, methanesulfonamide, tert. butanol/water, 0°C to rt, 72 h, 65%; j) dibutyltin(IV) oxide, trimethylamine,
p-toluenesulfonoyl chloride, in CH Cl , Ar, rt, 8 h, 90%; k) KOH, DMF, rt, 3h, 92%.
2
CO
3
, acetone, reflux, 4 h, 61%; b) 170 °C, Ar, neat, 40% (3-allylcatechol),
Cl , 0 °C to rt, 20 h, 56%; e) 8,
/THF, THF,
Cl , 0 °C,
2
2
3
, DMF, Ar, rt, 2.5 h, 93%; d) mCPBA, CH
2
2
methanol, rt or 60°C, 24-48 h, 37-59%; f) H
2
3
2
O
2
2
2
1
2
2
Binding affinities. The affinity profile of our novel agonist/antagonist hybrid ligands at β - and
1
β2-AR was determined in binding assays with membranes from HEK293T or CHO cells expressing either
murine β -AR or human β -AR together with the radioligand [³H]CGP12,177. The results from these
1
2
competition experiments (Table 1) were compared to carazolol and those of our previously described
27
catecholamines with different amino-substituents ((R)-1, (R)-2, Table 1) in order to identify the most
promising candidates for further pharmacological evaluation. All ligands sharing a 9H-carbazolyl head
group displayed substantial affinity for β-adrenergic receptors. Among the two enantiomers of carvedilol,
the (S)- isomer was found to be the eutomer, showing subnanomolar affinity for both receptor subtypes
(
K β -AR 0.21 nM, β -AR 0.03 nM) that was comparable to the affinity of racemic carazolol (K β -AR
i
1
2
i
1
1
2
ACS Paragon Plus Environment

Page 13 of 71
Journal of Medicinal Chemistry
1
2
3
0
.09 nM, β -AR 0.03 nM). Inversion of the stereocenter resulted in a 10- to 30-fold reduction of binding
2
affinity for the (R)-distomer (K 3.4 nM and 0.88 nM) and a complete removal of the hydroxyl group led
i
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
to a further loss of affinity (K 24 and 5.4 nM). Formal shortening the oxypropylene-spacer to the agonist-
i
type ethanolamine-scaffold resulted in overall lower but still nanomolar binding affinity. In this case, the
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
eutomer was (R)-15, (K 6.6 and 3.3 nM for β -and β -AR). Importantly, (R)-15 displays a spatial
i
1
2
orientation of the hydroxyl substituent comparable to (S)-carvedilol, since the removal of the
40
oxymethylene-unit in the linker results in a formal reversal of CIP-priorities . Again, inversion of the
hydroxyl-configuration ((S)-15) resulted in an about 50-fold loss in binding affinity for β -AR and β -
1
2
AR, whereas the removal of the hydroxyl function (10) mainly affected β -AR (K 330 nM) and β -AR
1
i
2
only to a small extent (K 26 nM). These findings are in good agreement with the important interactions
i
of the oxypropanol/ethanolamine moiety with Asp3.32 and Asn7.39 in the orthosteric binding site of β-
adrenergic receptors.5, 14
For ligands with a catecholamine head group, we have previously discovered that large lipophilic
10
amino-substituents derived from carvedilol or FAUC72 are well tolerated and may even lead to higher
2
7
affinity for β -AR and β -AR as compared to isoprenaline. We were therefore interested to determine
1
2
whether a similar trend would be observed after formal elongation of the spacer with an oxymethylene
group, the common feature of β-antagonists. Indeed, hybrid ligands (S)-22, (S)-26 and (S)-30 were able
to bind β -AR and β -AR with nanomolar affinity (K β -AR 6.5-14 nM, β -AR 7.2-21 nM), with the
1
2
i
1
2
highest affinities observed for (S)-30 comprising the largest appendage. Inversion of the absolute
configuration resulted in a 16- to 60-fold reduction of affinity for both receptor subtypes for (R)-22 and
(R)-30. Remarkably, the eudismic ratio was much higher for (S)-/(R)-26, as (R)-26 has a more than 400-
fold lower affinity (K β -AR 7,400 nM, β -AR 7,900 nM). However, the eudismic ratio is highly
i
1
2
dependent on the enantiomeric purity (enantiomeric excess) of the ligands, which slightly differs between
the six compounds. The presence of the large amino-substituent was generally confirmed to be beneficial
for β-adrenoceptor affinity, as a complete removal of the hydroxyl led to a dramatic loss in affinity for
1
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
4 and 28 (K > 3.8 µM), while 32 retained some affinity for β -AR (K 700 nM). In contrast to the well
i 2 i
tolerated insertion of an oxymethylene-unit, homologisation of the catecholamines by inserting only one
methylene unit dramatically impaired receptor recognition, especially at β -AR, where only micromolar
1
affinities were observed for (R)-/(S)-39 and the deshydroxy-analogue 43 (K ≥ 2.4 µM). At β -AR,
i
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
affinities were only slightly better and ranged from 740 nM to 5.2 µM, illustrating that elongation of the
distance between the catechol and the 2-aminoalcohol moiety by one carbon atom leads to compounds
that do not fit well to the shape of the orthosteric site of both, β - and β -AR. This may be explained by
1
2
a conformational change that is induced by the presence of the benzylic methylene protons. Compared
to catechols with an ethanolamine- or oxypropanolamine-spacer a significant steric repulsion of the
benzene ring is observed, leading to a different relative orientation of the phenolic hydroxyl groups and
the aminoalcohol moiety (Supplementary Figure S1). Interestingly, direct insertion of a methylene group
adjacent to the aromatic meta-hydroxyl is well tolerated, as demonstrated by the high β -AR affinity (K
2
i
0.25 nM) of the partial agonist salmeterol.
In general, ligands with the 9H-carbazole head group showed significantly higher affinity towards
β -AR and β -AR compared to the analogues comprising a catechol pharmacophore. This can be
1
2
explained by the crystal structures of carvedilol and carazolol bound to β -AR and β -AR, respectively,
1
2
which reveal an engagement of the carbazolyl-pharmacophore in an intensive lipophilic interaction
network that is not contacted by typical catechols.5, 9, 14, 15
1
4
ACS Paragon Plus Environment

Page 15 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 1. Affinities of the test compounds for murine  and human  adrenoceptor subtypes determined
by competition binding experiments with the radioligand [³H]CGP12,177.
1
2
K [nM]
i
Hybrid
Compound
Type Structure
R
-
m -AR
h -AR
1
2
carazolol
(racemic)
OH
H
N
0.09 ± 0.01 a
0.03 ± 0.01 a
-
HN
O
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
OH
H
N
(R)-
isoprenaline
HO
HO
-
-
-
-
220 ± 48 a,c
46 ± 10 a,c
OH
H
(±)-
N
550 ± 100 a
0.25 ± 0.07 a
HO
O
salmeterol
HO
O
O
OH
OH
H
N
HO
160 ± 130 b,c
85 ± 21 a,c
97 ± 33 b,c
18 ± 4 a,c
(
R)-1
-
-
-
-
HO
HO
HO
H
N
O
O
(R)-2
(
S)-carvedilol
R)-carvedilol
-
-
(S)-OH 0.21 ± 0.03 a
0.03 ± 0.01 a
O
O
OH
OH
R
H
N
(R)-OH 3.4 ± 0.3 a
0.88 ± 0.10 a
(
HN
O
R
(
S)-OH (R)-OH
3
-
H
H
24 ± 10 a
5.4 ± 0.3 a
26 ± 4b
10
I
330 ± 40 b
O
O
OH
OH
H
N
(R)-OH 6.6 ± 3.5b
3.3 ± 1.8b
(
(
(
R)-15
I
HN
(R)-OH (S)-OH
S)-15
S)-22
I
(S)-OH 340 ± 60b
170 ± 60b
(S)-OH 14 ± 5b
12 ± 1b
II
II
II
II
II
II
II
II
II
III
III
III
OH
OH
R
H
HO
HO
O
N
(R)-OH 860 ± 350b
420 ± 10b
(R)-22
(
S)-OH (R)-OH
2
4
H
>10,000b
(S)-OH 12 ± 4 a
6,900 ± 5,800b
(
S)-26
R)-26
21 ± 5 a
O
OH
OH
R
H
N
HO
HO
O
b
b
(
O
(R)-OH 7,400 ± 1,800 7,900 ± 2,000
(
S)-OH (R)-OH
>10,0000b
3,800 ± 800b
2
8
H
OH
OH
(S)-OH 6.5 ± 0.9 a
7.2 ± 0.6 a
120 ± 20 a
(S)-30
R)-30
R
H
HO
HO
O
N
O
O
(S)-OH (R)-OH (R)-OH 310 ± 120 a
(
3
2
H
1,200 ± 500b 700 ± 570b
(R)-OH >10,000b
5,200 ± 500b
740 ± 50b
(
R)-39
S)-39
OH
OH
HO
HO
O
O
R
H
N
(S)-OH 2,400 ± 100b
(
(R)-OH (S)-OH
43
H
>10,000b
920 ± 190 a
a
b
Values indicate mean ± SEM in nM derived from 3 to 10 individual experiments each performed in triplicate. Values
c
27
indicate mean ± SD derived from two individual experiments each performed in triplicate. Taken from ref.
ACS Paragon Plus Environment
15

Journal of Medicinal Chemistry
Page 16 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Receptor activation. To characterize the pharmacological behavior of our agonist/antagonist
hybrid compounds and to determine crucial structure-activity relationships, we examined a subset of our
ligands for their activation properties at β - and β -AR. It has been recently shown that distinct ligands
1
2
may lead to individual receptor conformations that may differentially engage the activation of G protein
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
and recruitment of β-arrestins. Thus, we chose two different intracellular interaction partners of β - and
1
β -AR: Gα , the canonical and best characterized signal transducer, and β-arrestin 2, an adaptor protein
2
S
that is recruited to ligand-activated receptors facilitating receptor sequestration/internalization and
signaling.41-44 Investigations of the two pathways were performed in HEK293T cells using
bioluminescence resonance energy transfer (BRET), either between membrane anchored GFP and
4
3, 45
luciferase-tagged β-arrestin 2 or luciferase-tagged Gα and GFP-tagged Gγ , respectively.
s
1
s 1
Figure 3. Dose-response curves for Gα protein activation and β-arrestin-2 recruitment at β -AR reveal differential activation
patterns for catecholamines (R)-1 and (R)-2 and agonist/antagonist hybrid ligands of type II. Data show the mean curve ±
s.e.m. of three to four independent experiments, each performed in duplicate.
ACS Paragon Plus Environment
16

Page 17 of 71
Journal of Medicinal Chemistry
1
2
3
In general, the observed signaling signatures for the ligands of type I and II were found to be
similar for the two investigated β-AR subtypes but revealed different activity profiles when G and -
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
s
arrestin-2 coupling was compared (Figures 3 and 4, Supplementary Figures S2 and S3, Table 2). Type
III ligands were not further investigated because of their lack in binding affinity.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2 s
Figure 4. At β -AR, dose-response curves for Gα protein activation and β-arrestin-2 recruitment confirm the agonist-like
properties of (R)-1 and (R)-2, while only low intrinsic activities were observed for type II-hybrid compounds. Data show the
mean curve ± s.e.m. of three to four independent experiments, each performed in duplicate.
The two ligands with prototypical catechol-ethanolamine substructure, (R)-1 and (R)-2, displayed
strong agonist activity for the activation of β - and β -AR, which is in good agreement with previous
1
2
studies on β -AR activation employing a second messenger assay for G protein activation and enzyme
2
2
7
fragment complementation technology for the recruitment of β-arrestin-2. (R)-1 was found to activate
the two investigated signaling pathways with higher potency as compared to (R)-2, in spite of its slightly
lower affinity. This effect was most pronounced for the activation of β -AR and might be explained by
1
ACS Paragon Plus Environment
17

Journal of Medicinal Chemistry
Page 18 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
the large appendage of (R)-2 extending to the vestibule site, a region were the two receptor subtypes are
1
6
less conserved and that is also addressed by the β -selective partial agonist salmeterol. Interestingly,
2
full agonism with a tendency towards a small super-agonistic activity (Emax 119-147%) compared to
isoprenaline was observed for the activation of Gα by (R)-1 and (R)-2 at both subtypes, β -AR and β -
s
1
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
AR. Slightly submaximal efficacy was observed for the β-arrestin-2 recruitment, pointing towards a
preferential activation of Gα over other signal transducers for these two ligands.
s
A direct comparison of ligands (R)-1 and (R)-15 (type I hybrid) clearly reveals that the overall
intrinsic activity of the ligands is determined by the nature of the aromatic head group. While
catecholamine (R)-1 potently activates the two investigated signal transducers at β -AR and β -AR, the
1
2
9H-carbazole (R)-15 was devoid of agonistic activity for the two pathways investigated, in spite of its
high binding affinity for both β-AR subtypes.
Interestingly, an elongation of the distance between the aromatic head group and the amino
alcohol function modulated ligand efficacy pathway specifically. Indeed, insertion of an oxymethylene
unit into (R)-1 resulting in (S)-26, substantially decreases β-arrestin-2 recruitment (Emax 22-28.3 % at β1-
AR in the presence or absence of GRKs 2 or 5), whereas maximal Gα activation is affected to a very
s
minor extent (Emax 116% at β -AR). The low intrinsic efficacy in the β-arrestin recruitment assay but also
1
the high ligand concentration required for the activation of Gα (EC 300-800 nM) suggest that type II
s
50
ligands also possess substantial antagonistic characteristics. Although activation of β -AR required
2
higher ligand concentrations compared to β -AR, a similar trend was observed with high intrinsic activity
1
detected for Gα , and only very weak recruitment of β-arrestin-2. These observations were also confirmed
s
for the long chain analog (S)-30 that was able to strongly activate Gα (E 118-150% for β - and β -
s
max
1
2
AR, respectively) and did not promote β-arrestin recruitment in the absence of GRKs and resulted in only
moderate β-arrestin-2 recruitment (Emax 17-42%) in presence of GRK2 or GRK5 at β - and β -AR,
1
2
respectively. The elongated isoprenaline (S)-22 was found to display the same characteristic activation
pattern but only possesses partial agonist properties for the activation of Gα (E 73% for β - and β -
s
max
1
2
ACS Paragon Plus Environment
18

Page 19 of 71
1
Journal of Medicinal Chemistry
AR). Overall, although GRKs expression moderately affected the recruitment of β-arrestin-2 for some of
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
the compounds, the preference for Gα over β-arrestin was not prevented by such expression. No
s
agonistic activity for Gα or β-arrestin was observed for carvedilol at β -AR and β -AR (Table 2).
s
1
2
Compared to the β -selective partial agonist salmeterol, the type II agonist/antagonist hybrid with the
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
isopropyl substituent (S)-22, displays a similar intrinsic activity for the activation of Gα , while ligands
s
with larger amino-substituents, (S)-26 and (S)-30, were found to possess significantly higher intrinsic
activity.
a
Table 2. β - and β -AR mediated activation of Gα and β-arrestin recruitment determined by BRET.
1
2
s
β
1
-AR
2
β -AR
β-arr-2
- GRK)
β-arr-2
(+ GRK2)
β-arr-2
(+ GRK5)
β-arr-2
(- GRK)
β-arr-2
(+ GRK2)
β-arr-2
(+ GRK5)
Gα
s
Gα
s
(
compound
Emax pEC50
Emax pEC50
Emax pEC50
Emax pEC50
Emax pEC50
Emax pEC50
Emax pEC50
Emax pEC50
100
± 5 ± 0.2
7.6 100 7.28 99.8
7.3 100 7.36
102
± 5 ± 0.2
8.2
100 7.34 99.6
± 2 ± 0.05 ± 2.7 ± 0.1
7.6
99 7.70
± 3 ± 0.08
(
R)-isoprenaline
± 2 ± 0.05 ± 2.1 ± 0.1
± 3 ± 0.07
7
±
9
5.3
8.1 6.7
± 0.5 ± 0.1
76 9.1
± 5 ± 0.3
127 7.2
± 8 ± 0.2
119 7.0
± 8 ± 0.2
24 9.9
± 1 ±0.2
90 6.57 89.9 6.8
± 1 ± 0.03 ±1.7 ± 0.1
84 6.14 83.3 6.3
± 2 ± 0.06 ± 2.2 ± 0.1
30 10.1
± 1 ± 0.2
(±)-salmeterol
< 5
-
< 5
-
< 5
-
5 ± 0.2
1
32 7.4
6 ± 0.1
47 5.7
± 16 ± 0.2
69
7.1 83.9
7.2
90 7.29
± 3 ± 0.08
107 7.00
± 2 ± 0.04
(
R)-1
R)-2
±
± 4 ± 0.2 ±3.4 ± 0.1
95 6.1 87.6 6.1
± 6 ± 0.1 ± 5.0 ± 0.1
1
91
6.3
98 6.38
± 2 ± 0.04
(
± 2 ± 0.1
(S)-carvedilol
R)-carvedilol
R)-15
(S)-22
< 5
< 5
< 5
-
-
< 5
< 5
-
-
-
< 5
< 5
< 5
-
-
< 5
< 5
-
-
< 5
< 5
-
-
< 5
< 5
< 5
< 5
< 5
< 5
-
-
-
-
-
-
< 5
< 5
-
-
< 5
< 5
-
-
(
(
-
< 5
21
-
< 5
22
± 2 ± 0.2
25 6.4
± 4 ± 0.3
42 5.1
-
< 5
73
± 12 ± 0.5
139 4.9
± 41 ± 0.7
118 4.9
± 40 ± 0.7
-
< 5
9.9
± 1.2 ± 0.3
12.3 6.1
± 2.4 ± 0.4
-
< 5
18
± 1 ± 0.1
11 6.2
± 2 ± 0.4
17 5.7
± 1 ± 0.1
-
7
3
6.2
6.2 19.7
6.0
6.5
6.1
6.5
6.6
±
10 ± 0.3
116 6.1
± 11 ± 0.1
50 5.6
18 ± 0.2 ± 33
± 3 ± 0.2 ± 3.2 ± 0.3
22 6.5 28.3 6.4
± 4 ± 0.4 ± 2.7 ± 0.2
32 33.3 5.6
± 1 ± 3.4 ± 0.2 ± 19 ± 0.5
(
S)-26
S)-30
1
±
5
(
< 5
-
a
Data represent mean ± s.e.m. of three to four independent experiments, each performed in duplicate. Emax is given relative
[%] to the maximum response of isoprenaline.
Molecular dynamics. In light of the recently published co-crystal structure of β -AR in complex
2
1
6
with a specific nanobody and the G protein-biased partial agonist salmeterol, we were interested to
ACS Paragon Plus Environment
19

Journal of Medicinal Chemistry
Page 20 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
obtain further insights into the origins of the functionally selective agonist activity observed for our type
II catechol-beta blocker hybrid ligands. This is of particular interest, as very early studies on these
catecholoxypropanolamines suggested bent conformations to superimpose with typical catecolamines
and thus causing the agonistic behavior observed with isolated arterial and tracheal tissue long before the
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
8
first structural information on GPCRs was available. Although being far from proof, docking and long-
term molecular dynamics simulations should be able to provide reasonable hypotheses on the binding
mode and key interactions between the receptor and this class of agonist/antagonist hybrid molecules.
We chose (S)-22, as representative candidate for our in silico investigations, as this ligand possesses high
similarity to the previously crystallized endogenous agonist adrenaline and the high affinity antagonist
carazolol. Moreover, the small isopropyl side chain is less likely to cause fluctuations during the
simulations eventually caused by initial misplacement during docking. We performed molecular
dynamics simulations at β -AR since high-resolution structural information on the distinct receptor
2
states9, 14, 16 is available. As reference systems, we used epinephrine at the β -AR in the fully active state
2
(
PDB entry 4LDO) and β -AR in complex with the G protein-biased partial agonist salmeterol (PDB
2
entry 6MXT). As (S)-22 showed G protein-biased agonism in our functional assays, we performed our
simulations with the salmeterol-induced β -AR structure (6MXT). To rule out a possible influence on
2
the binding mode provoked by the different receptor states, we also simulated the three ligands in the
second receptor structure as control experiments. Each simulation was repeated six times, with a
simulation time of 2 µs, each. In all simulation conditions, the three ligands maintained a stable binding
pose, as we observed a stable salt bridge to Asp3.32 and low root-mean-square deviation (rmsd) values
for the heavy atoms of the ligands (Figure 5a, Supplementary Figure S4). A comparison of the
simulations of epinephrine and (S)-22 reveals differences in the hydrogen bond network established
between the ligands and crucial residues within the orthosteric binding pocket. For epinephrine, the
hydrogen bond between the β-OH group and the sidechain of Asn7.39, which is present in the crystal
1
4
structure, was only observed in roughly 20% of the simulation time. In contrast, this interaction was
ACS Paragon Plus Environment
20

Page 21 of 71
1
Journal of Medicinal Chemistry
observed in about 55% of the simulation time for (S)-22 (Figure 5d). This is probably a result of a better
orientation of the hydroxyl group towards Asn7.39, permitted by the increased distance of the β-OH to the
aromatic head group of (S)-22. For salmeterol, no significant differences in the frequency of this H-bond
interaction were observed (Supplementary Figure S5).
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 5. (a) Schematic 2D representation of the key interaction motifs of epinephrine and (S)-22 within the orthosteric
2
binding pocket of β -AR. (b/c) Representative frames of the most frequent binding pose of epinephrine (b) and (S)-22 (c) in
the β -AR. The distances of interest are shown with their scalar value in Å. (d) Fraction of simulation time in which the
2
distances for the described interactions are below 2.0 Å (measurement from H-bond donor hydrogen to H-bond acceptor
heavy atom). Data show mean ± s.e.m. of six independent simulation runs (grey points). Statistical significance was assessed
by one-sided Welch’s unequal variance t test (* p < 0.05, ** p < 0.01, *** p < 0.001, NS; not significant). (e) The two dihedral
angles of Ser5.46 χ and χ are plotted against each other. Data shown for one representative simulation run for the two systems.
1
2
Schematic representation of Ser5.46 indicating the used nomenclature of dihedrals.
ACS Paragon Plus Environment
21

Journal of Medicinal Chemistry
Page 22 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
For the hydrogen bonds to the crucial serines in TM5, we found a slightly lower prevalence of
the H-bond between the para-hydroxyl and Ser5.46 for (S)-22 compared to epinephrine, whereas we did
not observe any differences for the meta-hydroxyl H-bond to Ser5.42 (Figure 5d). In contrast, the
interaction with Ser5.42 was more frequently observed for β -AR in complex with salmeterol, which is in
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
6
good agreement with previously published data (Supplementary Figure S5). In the so far published
active state crystal structures of β -AR, Ser is rotated towards TM3 and interacts with Thr3.37 (Figure
5
.46
2
1
), whereas it faces to the opposite side in inactive crystal structures. The slightly lower prevalence of
the H-bond to Ser5.46 for (S)-22 could influence the conformation of Ser5.46 and explain the observed,
more diverse conformations of Ser5.46 compared to epinephrine (Figure 5b,e). This behavior is more
pronounced in the simulations of (S)-22 at the salmeterol-induced simulation system than the
epinephrine-induced simulation system (Supplementary Figure S6). This suggests that the observed
difference in conformations of Ser5.46 could result from the state of the receptor and is not necessarily
provoked by the ligand itself. However, different rotamers of Ser5.46 are much less frequently observed
when the full agonist epinephrine is simulated in the salmeterol-based receptor structure.
Compared to epinephrine, (S)-22 also shows a reduced H-bond formation of the meta-hydroxyl to the
sidechain of Asn6.55, similar to the recently reported data on salmeterol which our simulations were able
to reproduce (Figure 5d, Supplementary Figure S5). Both (S)-22 and salmeterol show reduced interaction
with Asn6.55 compared to epinephrine and a lower efficacy (~75%), suggesting that this interaction seems
to influence the extent by which a ligand can activate Gα -coupling at the β -AR. In agreement with these
s
2
findings, the β -AR partial agonist xamoterol, and other derivatives which are lacking the meta-hydroxyl
1
group and thus unable to establish an H-bond to Asn6.55, have no or little activity at β -AR.
46, 47
The
2
observed submaximal intrinsic activity together with apparent preference for G protein activation over
β-arrestin-2 recruitment of (S)-22 suggest the interaction to Asn6.55 to be crucial but most likely not the
sole determinant for β -AR activation and ligand bias. Importantly, the homologous residue (His6.55) has
2
ACS Paragon Plus Environment
22

Page 23 of 71
1
Journal of Medicinal Chemistry
previously been implicated in biased signaling and partial agonist activity at another aminergic GPCR,
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
4
8, 49
the dopamine D receptor.
2
CONCLUSION
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Starting from prototypical adrenoceptor ligands, catecholamine type agonists and carbazolyl-
containing beta-blockers, we designed synthesized and pharmacologically characterized three different
chemotypes of agonist/antagonist hybrids. All ligands were synthesized as isolated enantiomers, enabling
the identification of the more active stereoisomer (eutomer) by means of radioligand binding studies.
Further investigations of ligand-mediated receptor activation using BRET-biosensors revealed a
predominant effect of the aromatic head group on the overall intrinsic activity of our ligands, as ligands
with a 9H-carbazolyl head group were antagonists irrespective of the employed linker region. Type II
hybrid ligands being composed of a catechol head group and an oxypropylene-spacer were found to
possess significant intrinsic activity for the activation of Gα , while they only showed weak or even no
s
intrinsic activity for the recruitment of β-arrestin-2 at both, β - and β -AR. Results from our unbiased
1
2
molecular dynamics simulations suggest that the difference in G protein efficacy and β-arrestin
recruitment of the agonist/antagonist hybrid (S)-22, the full agonist epinephrine and the β -selective, G
2
protein-biased partial agonist salmeterol depends on the ligands’ ability to form and maintain specific
hydrogen bonds between Ser5.46 and Asn6.55 and the para-hydroxy and meta-substituent of the ligands,
respectively.
EXPERIMENTAL SECTION
Chemistry. Reagents and dry solvents were of commercial grade and used as purchased. HPLC-
MS was run on a BRUKER ESQUIRE 2000 ion-trap mass spectrometer or on a BRUKER amaZon SL
mass spectrometer using ESI or APCI as ionization source. HRMS-ESI was performed on a Bruker
Daltonic microTOF II focus TOF-MS spectrometer using ESI as ionization source. HRMS-EI was run
ACS Paragon Plus Environment
23

Journal of Medicinal Chemistry
Page 24 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
on a JEOL JMS-GC mate II with a solid inlet and EI (70 eV) using Peak-Matching (M/ΔM > 5000).
NMR spectra were obtained on a Bruker Avance 360, Bruker Avance 400 or a Bruker Avance 600
1
13
spectrometer at 300 K. H and C chemical shifts are given in ppm (δ) relative to TMS in the solvents
indicated. IR spectra were performed on a Jasco FT/IR 4100 spectrometer using a film of substance on a
NaCl plate. Purification by column chromatography was conducted using silica gel 60 (40-63 µm mesh,
Merck) and eluents as binary mixtures with the volume ratios and modifiers indicated. Preparative HPLC
was performed on an Agilent 1100 preparative series HPLC system combined with a VWD detector. As
HPLC column, a MACHEREY-NAGEL Varioprep VP 250/10 Nucleodur C18 HTec (10 x 250 mm, 5
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
µm) was used (flow rate 4 mL/min, λ = 254 nm). TLC analyses were performed using Merck 60 F254
aluminum sheets and analyzed by UV light (254 nm) and ninhydrin spray reagent for amines. Analytical
HPLC was performed on an AGILENT 1200 series HPLC system employing a DAD detector. As HPLC
column, a ZORBAX ECLIPSE XDB-C8 (4.6 x 150 mm, 5 µm) was used. HPLC purity was measured
using the following binary solvent systems: 1) 10% methanol in 0.1% aq. formic acid for 3 min, from
1
1
0% to 100% methanol in 15 min, 100% methanol for 6 min, from 100% to 10% methanol in 3 min and
0% methanol for 3 min. 2) 10% methanol in 0.1% aq. trifluoroacetic acid for 3 min, from 10% to 100%
methanol in 15 min, 100% methanol for 6 min, from 100% to 10% methanol in 3 min and 10% methanol
for 3 min. 3) 10 % acetonitrile in 0.1% formic acid for 3 min, from 10% to 100% acetonitrile in 15 min,
1
5
00 % acetonitrile for 6 min, from 100% to 10 % acetonitrile in 3 min, 10 % acetonitrile for 3 min. 4)
-80 % acetonitrile in 0.1% trifluoroacetic acid for 18 min, 80% to 95 % acetonitrile in 2 min, 95 %
acetonitrile for 2 min, from 95% to 5 % acetonitrile in 3 min, 5 % acetonitrile for 3 min. For all systems,
a flow rate of 0.5 mL/min and a detection wavelength of λ = 220 nm, 254 nm or 280 nm was employed.
The purity of all test compounds was determined to be >95%. Chiral analytical HPLC was performed on
an AGILENT 1100 series HPLC system employing a VWD detector (λ = 254 nm). As chiral HPLC
column, a DAICEL Chiralpak IB or IC (4.6 x 250 mm, 5µm) was used. Chiral purity was measured using
the following binary solvent systems: C1) 15% isopropanol + 0.1% diethylamine in acetonitrile, isocratic,
ACS Paragon Plus Environment
24

Page 25 of 71
Journal of Medicinal Chemistry
1
2
3
0
.7 mL/min; C2) 70 % isopropanol + 0.2% diethylamine in hexane, isocratic, 0.7 mL/min; C3) 10%
isopropanol + 0.1% diethylamine in acetonitrile, isocratic, 0.7 mL/min; C4) 30% isopropanol + 0.2%
diethylamine in hexane, isocratic, 0.7 mL/min. Chiral purity was determined to be >95%. Optical rotation
values were obtained from a Jasco P2000 polarimeter with the solvents indicated.
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(R)-1-[(9H-Carbazol-4-yl)oxy]-3-{[2-(2-methoxyphenoxy)ethyl]amino}propan-2-ol
((R)-
carvedilol). A solution of (R)-45 (33 mg, 0.137 mmol) and 8 (46 mg, 0.274 mmol) in isopropanol (7
mL) was heated to reflux for 4 h. After completion of the reaction, the solvent was evaporated and the
residue was purified by column chromatography (CH Cl /methanol 95:5 + 0.2% NH ) to yield a yellow
2
2
3
oil. The product was taken up in water (20 mL) and lyophilized to give (R)-carvedilol as a white powder
(37 mg, 66%). IR (NaCl): 3400, 3062, 2931, 2837, 1895, 1591, 1506, 1455, 1334, 1253, 1100, 1025, 746
-1
1
cm . H NMR (CDCl , 600 MHz, δ): 2.62 (s(b), 2 H), 3.00 (dd, J = 12.5, 7.9 Hz, 1 H), 3.10 (dd, J =
3
1
2.5, 3.3 Hz, 1 H), 3.10 – 3.15 (m, 2 H), 3.80 (s, 3 H), 4.16 (t, J = 5.1 Hz, 2 H), 4.19 (dd, J = 9.3, 5.1 Hz,
1 H), 4.26 (dd, J = 9.3, 5.5 Hz, 1 H), 4.37 – 4.42 (m, 1 H), 6.64 (d, J = 7.9 Hz, 1 H), 6.84 – 6.91 (m, 3 H),
6
.91 – 6.95 (m, 1 H), 7.03 (d, J = 7.9 Hz, 1 H), 7.18 (ddd, J = 7.9, 7.1, 1.0 Hz, 1 H), 7.29 (dd, J = 7.9,
.9 Hz, 1 H), 7.35 (ddd, J = 7.6, 7.1, 0.9 Hz, 1 H), 7.38 (d, J = 7.6 Hz, 1 H), 8.18 (s, 1 H), 8.24 (d, J =
7
1
3
7.9 Hz, 1 H). C NMR (CDCl , 90 MHz, δ): 48.5, 51.9, 55.8, 67.9, 68.2, 70.1, 101.3, 103.9, 110.0, 112.0,
3
1
12.7, 114.7, 119.7, 121.0, 122.1, 122.5, 122.9, 125.0, 126.7, 138.7, 141.0, 147.9, 149.8, 155.0. ESI-MS:
+
+
20
m/z 407.5 [M+H] . HRMS-EI: [M ] calcd for C H N O , 406.1893; found 406.1892. [α] +14.5 (c
2
4
26
2
4
D
0
.4, glacial acetic acid). HPLC (system 1): t = 16.1 min, purity 99%; (system 3): t = 13.6 min, purity
R R
2
0
99%. Chiral HPLC (system c1): t = 9.3 min, ee 96%. (S)-carvedilol (38 mg, 70%, [α] -14.7 (c 0.4,
R
D
glacial acetic acid) was prepared from (S)-45 (32 mg, 0.133 mmol) and 8 (45 mg, 0.267 mmol) described
above for (R)-carvedilol. HPLC (system 1): t = 16.2 min, purity 99%; (system 3): t = 13.5 min, purity
R
R
99%. Chiral HPLC (system c1): t = 10.0 min, ee 95%.
R
3
-[(9H-Carbazol-4-yl)oxy]-N-[2-(2-methoxyphenoxy)ethyl]propan-1-amine (3). A solution
of 44 (50 mg, 0.16 mmol), 8 (69mg, 0.41 mmol) and KI (41 mg, 0.25 mmol) in CH CN (2.5 mL) was
3
ACS Paragon Plus Environment
25

Journal of Medicinal Chemistry
Page 26 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
refluxed under a nitrogen atmosphere for 5 h. After the reaction has cooled to room temperature, water
was added and the aqueous phase was extracted with CH Cl . The combined organic layers were washed
2
2
with brine, dried over Na SO and evaporated. The crude product was purified by column
2
4
chromatography (CH Cl /methanol 96:4 + 0.2% NH ) to give 3 as a yellow oil which solidified upon
2
2
3
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
standing (47 mg, 73%). IR (NaCl): 3407, 3064, 2943, 1604, 1508, 1457, 1331, 1254, 1224, 1128, 1098,
-1 1
1027, 743 cm . H NMR (CDCl , 600 MHz, δ): 2.09 (s(b), 2 H), 2.13 – 2.2 (m, 2 H), 3.0 (t, J = 7 Hz,
3
2
H), 3.07 (t, J = 5.4 Hz, 2 H), 3.79 (s, 3 H), 4.13 (t, J = 5.4 Hz, 2 H), 4.27 (t, J = 6.1 Hz, 2 H), 6.64 (d,
J = 7.9 Hz, 1 H), 6.83 – 6.94 (m, 4 H), 7.00 (d, J = 7.9 Hz, 1 H), 7.19 (ddd, J = 7.8, 6.2, 1.8 Hz, 1 H),
1
3
7
.29 (dd, J = 7.9, 7.9 Hz, 1 H), 7.33 – 7.37 (m, 2 H), 8.21 (s, 1 H), 8.29 (d, J = 7.8 Hz, 1 H). C NMR
(CDCl , 150 MHz, δ): 29.9, 46.9, 48.9, 55.7, 66.1, 68.6, 101.0, 103.4, 109.9, 111.8, 112.6, 114.1, 119.6,
3
20.9, 121.5, 122.7, 123.0, 124.8, 126.6, 138.7, 140.9, 148.3, 149.6, 155.5. ESI-MS: m/z 391.1 [M+H]+.
1
+
HRMS-EI: [M ] calcd for C H N O , 390.1944; found 390.1942. HPLC (system 2): t = 17.6 min,
2
4
26
2
3
R
purity 98%; (system 4): t = 13.8 min, purity 97%.
R
9
H-Carbazol-4-yl trifluoromethanesulfonate (5). A solution of 4 (1422 mg, 7.76 mmol) in
CH Cl (30 mL) was precooled to -10 °C under argon atmosphere. Triethylamine (1.08 mL, 7.76 mmol)
2
2
was added and the mixture was stirred for 15 minutes before trifluoromethanesulfonic anhydride (1.57
mL, 9.31 mmol) was added dropwise. The reaction was allowed to warm to room temperature and stirred
for further 60 minutes. The reaction was terminated by the addition of water and the aqueous phase was
extracted with CH Cl . The combined organic layers were washed with brine, dried over Na SO and
2
2
2
4
evaporated. The crude product was purified by column chromatography (hexane/ethyl acetate 9:1 – 4:1)
to give 5 as a grey oil which solidified upon standing (2066 mg, 84%). IR (NaCl): 3406, 3067, 2916,
-1 1
1
614, 1457, 1419, 1329, 1210, 1140, 1008, 850, 750, 605 cm . H NMR (CDCl , 360 MHz, δ): 7.19 (dd,
3
J = 4.4, 4.4 Hz, 1 H), 7.33 (ddd, J = 8, 6.7, 1.4 Hz, 1 H), 7.39 – 7.42 (m, 2 H), 7.43 – 7.47 (m, 1 H), 7.47
1
3
–
7.53 (m, 1 H), 8.25 (s, 1 H), 8.30 (d, J = 8 Hz, 1 H). C NMR (CDCl , 90 MHz, δ): 110.6, 110.8,
3
111.3, 115.9, 117.0, 119.9, 120.6, 122.7, 126.1, 127.1, 139.4, 141.4, 144.5. ESI-MS: m/z 316.4 [M+H]+.
26
ACS Paragon Plus Environment

Page 27 of 71
Journal of Medicinal Chemistry
1
2
3
(
Z)-4-(2-Ethoxyvinyl)-9H-carbazole (6). To a suspension of 5 (500 mg, 1.59 mmol),
bis(triphenylphosphine)palladium(II) dichloride (56 mg, 0.08 mmol) and lithium chloride (538 mg, 12.7
mmol) in DMF (5 mL) was added Z-tributyl(2-ethoxyethenyl)stannane (0.53 mL; 1.59 mmol). The
mixture was stirred at 90°C under argon atmosphere for 3 h. After cooling to room temperature, ethyl
acetate was added and the resulting suspension was filtered. Water was added to the filtrate and the
aqueous phase was extracted with ethyl acetate. The combined organic layers were washed with brine,
dried over Na SO and evaporated. The residue was taken up in CH CN (20 mL) and the dark suspension
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
4
3
was filtered again. The filtrate was evaporated and the crude product subsequently purified by column
chromatography (hexane/ethyl acetate 9:1) to give 6 as an orange oil which solidified upon standing (375
-1
mg, 90%). IR (NaCl): 3407, 3054, 2979, 1642, 1457, 1432, 1386, 1325, 1252, 1100, 999,801, 725 cm .
1
H NMR (CDCl , 360 MHz, δ): 1.39 (t, J = 7.1 Hz, 3 H), 4.05 (q, J = 7.1 Hz, 2 H), 6.17 (d, J = 7.2 Hz,
3
1
H), 6.50 (d, J = 7.2 Hz, 1 H), 7.20 – 7.27 (m, 2 H), 7.36 – 7.44 (m, 3 H), 7.83 (d, J = 7.6 Hz, 1 H), 8.02
1
3
(s, 1 H), 8.27 (d, J = 7.9 Hz, 1 H). C NMR (CDCl , 150 MHz, δ): 15.4, 69.1, 103.0, 108.1, 110.3, 119.2,
3
20.1, 120.5, 122.8, 123.7, 125.0, 125.6, 131.1, 139.6, 139.8, 147.5. ESI-MS: m/z 238.4 [M+H]+.
1
2-(9H-Carbazol-4-yl)acetaldehyde (7). To a solution of 6 (100 mg, 0.42 mmol) in acetone (2
mL) was added 2 M HCl (0.42 mL, 0.84 mmol) and the mixture was heated to reflux for 30 minutes.
After the reaction has cooled to room temperature, the solvent was evaporated in vacuo. The residue was
taken up in CH Cl , water was added and the aqueous phase was extracted with CH Cl . The combined
2
2
2
2
organic layers were washed with brine, dried over Na SO and evaporated to give 7 as a light brown oil.
2
4
The crude product was used for the next reaction without further purification.
-(2-Methoxyphenoxy)ethan-1-amine (8). A solution of 9 (1000 mg, 4.78 mmol) in 5 M HCl
2.75 mL) was heated to reflux for 21 h. After cooling to room temperature, the reaction mixture was
adjusted to pH 9 with 5 M NaOH and extracted with CH Cl . The combined organic layers were washed
2
(
2
2
with brine, dried over Na SO and evaporated to give 8 as a yellow oil (801 mg, 100%). IR (NaCl): 3407,
2
4
-1 1
3
066, 2942, 1593, 1506, 1456, 1328, 1254, 1224, 1180, 1124, 1023, 745 cm . H NMR (CDCl , 360
3
ACS Paragon Plus Environment
27

Journal of Medicinal Chemistry
Page 28 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
MHz, δ): 1.65 (s(b), 2 H), 3.02 – 3.19 (m, 2 H), 3.86 (s, 3 H), 4.05 (t, J = 5.3 Hz, 2 H), 6.86 – 6.97 (m,
1
3
4
H). C NMR (CDCl , 90 MHz, δ): 41.5, 55.9, 71.6, 112.0, 114.1, 120.9, 121.5, 148.4, 149.7. ESI-MS:
3
m/z 168.2 [M+H]+.
N-[2-(2-Methoxyphenoxy)ethyl]acetamide (9). A mixture of 2-methoxyphenol (1.1 mL, 10
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
mmol) and 2-methyl-2-oxazoline (0.84 mL, 10 mmol) was heated in a sealed tube to 160°C for 40 h
under argon atmosphere. After the reaction has cooled to room temperature, 1 M NaOH solution was
added and the aqueous phase was extracted with ethyl acetate. The combined organic layers were washed
with 1 M NaOH and brine, dried over Na SO and evaporated. The crude product was recrystallized from
2
4
hexane/ethyl acetate to obtain 9 as light yellow needles (1193 mg, 57%). IR (NaCl): 3544, 3276, 3067,
-1 1
2
939, 2878, 2837, 1652, 1506, 1374, 1253, 1180, 1125, 1026, 912, 747 cm . H NMR (CDCl , 600 MHz,
3
δ): 2.00 (s, 3 H), 3.63 – 3.66 (m, 2 H), 3.87 (s, 3 H), 4.08 (t, J = 5.0 Hz, 2 H), 6.24 (s, 1 H), 6.88 – 6.93
13
(
m, 3 H), 6.95 – 6.99 (m, 1 H). C NMR (CDCl , 150 MHz, δ): 23.3, 39.0, 55.8, 68.9, 111.9, 115.0,
3
1
21.1, 122.2, 147.8, 149.8, 170.1. ESI-MS: m/z 232.4.0 [M+Na]+.
N-[2-(9H-Carbazol-4-yl)ethyl]-2-(2-methoxyphenoxy)ethan-1-amine (10). To a solution of 8
(99 mg, 0.59 mmol) and sodium triacetoxyborohydride (536 mg, 2.53 mmol) in THF (7 mL) was added
a solution of crude 10 (~0.4 mmol) in THF (7 mL). The reaction was stirred at room temperature for 2.5
h under argon atmosphere. The mixture was filtered over celite and the solvent was evaporated under
vacuum. The residue was taken up in CH Cl , washed with saturated NaHCO solution and brine, dried
2
2
3
over Na SO and evaporated. The crude product was purified by column chromatography
2
4
(CH Cl /methanol 99:1 – 96:4 + 0.2% NH ) to give 10 as a light yellow oil which solidified upon standing
2
2
3
(55 mg, 36% over two steps). IR (NaCl): 3437, 2953, 1676, 1506, 1458, 1385, 1327, 1254, 1204, 1126,
-1 1
1
024, 801, 668 cm . H NMR (CDCl , 360 MHz, δ): 3.11 (t, J = 5.4 Hz, 2 H), 3.18 (t, J = 7.6 Hz, 2 H),
3
3.46 (t, J = 7.6 Hz, 2 H), 3.83 (s, 3 H), 4.15 (t, J = 5.4 Hz, 2 H), 6.84 – 6.97 (m, 4 H), 7.05 (dd, J = 6.5,
.7 Hz, 1 H), 7.22 (ddd, J = 8.0, 6.3, 1.7 Hz, 1 H), 7.28 – 7.47 (m, 4 H), 8.13 (d, J = 8.0 Hz, 1 H), 8.22
1
1
3
(
s, 1 H). C NMR (CDCl , 90 MHz, δ): 34.9, 48.8, 49.3, 55.9, 68.8, 108.8, 110.5, 112.0, 114.4, 119.6,
3
ACS Paragon Plus Environment
28

Page 29 of 71
Journal of Medicinal Chemistry
1
2
3
1
20.5, 121.0, 121.4, 121.6, 122.5, 123.1, 125.3, 125.8, 135.2, 139.5, 139.9, 148.4, 149.8. ESI-MS: m/z
+
+
3
61.7 [M+H] . HRMS-ESI: [M+H] calcd for C H N O , 361.19105; found 361.19124. HPLC (system
23 25 2 2
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1): t = 17.6 min, purity 99%; (system 3): t = 15.3 min, purity 97%.
R
R
4
-Vinyl-9H-carbazole (11). To
a
suspension of
5
(1877 mg, 5.96 mmol),
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
bis(triphenylphosphine)palladium(II) dichloride (209 mg, 0.30 mmol) and lithium chloride (2.0 g, 47.4
mmol) in DMF (15 mL) was added Z-tributyl(ethenyl)stannane (1.74 mL; 5.96 mmol). The mixture was
stirred at 90°C under argon atmosphere for 2 h. After cooling to room temperature, water was added and
the aqueous phase was extracted with ethyl acetate. The combined organic layers were washed with
brine, dried over Na SO and evaporated. The residue was taken up in CH CN (20 mL) and the dark
2
4
3
suspension was filtered. The filtrate was evaporated and the crude product subsequently purified by
column chromatography (hexane/ethyl acetate 9:1 – 8:2) to give 11 as a light orange oil which solidified
upon standing (1046 mg, 87%). IR (NaCl): 3587, 3054, 2955, 1569, 1457, 1382, 1324, 1256, 1151, 985,
-1 1
9
1
10798, 726 cm . H NMR (CDCl , 360 MHz, δ): 5.51 (dd, J = 10.9, 1.5 Hz, 1 H), 5.88 (dd, J = 17.4,
3
.5 Hz, 1 H), 7.20 – 7.26 (m, 1 H), 7.30 – 7.43 (m, 5 H), 7.69 (dd, J = 17.4, 10.9 Hz, 1 H), 8.03 (s, 1 H),
1
3
8.21 (d, J = 8 Hz, 1 H). C NMR (CDCl , 90 MHz, δ): 109.7, 110.5, 116.4, 117.2, 119.6, 120.6, 123.0,
3
23.3, 125.6, 125.8, 134.3, 135.3, 139.7. APCI-MS: m/z 194.2 [M+H]+.
1
(
R)-1-(9H-Carbazol-4-yl)ethane-1,2-diol ((R)-12). AD-mix β (2.92 g) was added portionwise
to a solution of 11 (400 mg, 2.07 mmol) in a tert-butanol/water mixture (1:1, 20 mL) at 0°C. The resulting
suspension was stirred at 0°C for 4 h. Na SO (3.3 g) was added and the mixture stirred for further 15
2
3
minutes. Water was then added and the aqueous phase was extracted with CH Cl . The combined organic
2
2
layers were washed with brine, dried over Na SO and evaporated. The crude product was purified with
2
4
column chromatography (CH Cl /methanol 99:1) to give (R)-12 as a light brown solid (393 mg, 84%).
2
2
-1 1
IR (NaCl): 3409, 3052, 2934, 1605, 1458, 1439, 1327, 1256, 1099, 1058, 1024, 867, 728 cm . H NMR
(
CD OD, 600 MHz, δ): 3.70 (dd, J = 11.6, 8.6 Hz, 1 H), 3.97 (dd, J = 11.6, 3 Hz, 1 H), 5.71 (dd, J = 8.6,
3
3
Hz, 1 H), 7.18 (ddd, J = 7.9, 7, 0.9 Hz, 1 H), 7.34 – 7.40 (m, 4 H), 7.45 – 7.48 (m, 1 H), 8.16 (d, J =
29
ACS Paragon Plus Environment