Synthesis and inhibitory evaluation of 3-linked imipramines for the exploration of the S2 site of the human serotonin transporter

Article abstract of DOI:10.1016/j.bmc.2016.04.039

The human serotonin transporter is the primary target of several antidepressant drugs, and the importance of a primary, high affinity binding site (S1) for antidepressant binding is well documented. The existence of a lower affinity, secondary binding site (S2) has, however, been debated. Herein we report the synthesis of 3-position coupled imipramine ligands from clomipramine using a copper free Sonogashira reaction. Ligand design was inspired by results from docking and steered molecular dynamics simulations, and the ligands were utilized in a structure-activity relationship study of the positional relationship between the S1 and S2 sites. The computer simulations suggested that the S2 site does indeed exist although with lower affinity for imipramine than observed within the S1 site. Additionally, it was possible to dock the 3-linked imipramine analogs into positions which occupy the S1 and the S2 site simultaneously. The structure activity relationship study showed that the shortest ligands were the most potent, and mutations enlarging the proposed S2 site were found to affect the larger ligands positively, while the smaller ligands were mostly unaffected.

Full text of DOI:10.1016/j.bmc.2016.04.039

Bioorganic & Medicinal Chemistry 24 (2016) 2725–2738
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Synthesis and inhibitory evaluation of 3-linked imipramines for the
exploration of the S2 site of the human serotonin transporter
a,y
a,c
b,y
b,
a,c,y,à
a
b
Anne Brinkø , Maja T. Larsen , Heidi Koldsø
, Louise Besenbacher , Anders Kolind ,
⇑
a,
⇑
Birgit Schiøtt , Steffen Sinning , Henrik H. Jensen
a
Department of Chemistry, Aarhus University, Denmark
Laboratory of Molecular Neurobiology, Translational Neuropsychiatry Unit, Department of Clinical Medicine, Aarhus University, Risskov, Denmark
inSPIN & iNANO Centres, Langelandsgade 140, 8000 Aarhus C, Denmark
b
c
a r t i c l e i n f o
a b s t r a c t
Article history:
The human serotonin transporter is the primary target of several antidepressant drugs, and the impor-
tance of a primary, high affinity binding site (S1) for antidepressant binding is well documented. The exis-
tence of a lower affinity, secondary binding site (S2) has, however, been debated. Herein we report the
Received 19 November 2015
Revised 16 April 2016
Accepted 20 April 2016
Available online 21 April 2016
synthesis of 3-position coupled imipramine ligands from clomipramine using
a copper free
Sonogashira reaction. Ligand design was inspired by results from docking and steered molecular dynam-
ics simulations, and the ligands were utilized in a structure–activity relationship study of the positional
relationship between the S1 and S2 sites. The computer simulations suggested that the S2 site does
indeed exist although with lower affinity for imipramine than observed within the S1 site.
Additionally, it was possible to dock the 3-linked imipramine analogs into positions which occupy the
S1 and the S2 site simultaneously. The structure activity relationship study showed that the shortest
ligands were the most potent, and mutations enlarging the proposed S2 site were found to affect the
larger ligands positively, while the smaller ligands were mostly unaffected.
Keywords:
Antidepressants
Sonogashira coupling
Bivalent ligands
Serotonin
Transporter
Depression
Ó 2016 Elsevier Ltd. All rights reserved.
1
. Introduction
The human serotonin transporter (hSERT) belongs to the Neuro-
Even though pharmacological treatment of depression has
come a long way, it still suffers from several challenges, such as
latency periods of 3–4 weeks, a range of severe side effects, and a
4
transmitter: Sodium Symporters (NSS) family and it regulates the
synaptic concentration of serotonin in the brain by reuptake of
serotonin. The transporter thus plays a critical role in modulat-
ing important physiological functions such as mood, appetite,
high percentage of non-responding patients. As a result, research
into the fundamental underlying mechanism of hSERT and its
1
,2
2,5
regulation remains an important and active area of research. Res-
olution of a crystal structure of the bacterial (Aquifex aeolicus) amino
2
6
and sleep. hSERT is the primary target of several antidepressant
acid transporter LeuTAa has made it possible to construct a homol-
7
,8
drugs belonging to the class of selective serotonin reuptake inhibi-
tors (SSRIs) such as citalopram and sertraline, but also less selec-
tive drugs from the tricyclic antidepressants class (TCAs) such as
clomipramine and imipramine.2 These drugs inhibit hSERT regu-
lated reuptake of serotonin from the synaptic cleft, effectively rais-
ing the concentration of serotonin available to the brain, thus
relieving the symptoms of depression through a still unknown
ogy model of hSERT with the primary (S1) binding pocket occu-
9
,10
10,11
12
13
pied by imipramine,
phenylpiperazine,
escitalopram,
mazindol, fluoxetine,
1
4
15
noribogaine15 and serotonin.
16,17
cocaine,
The TCA clomipramine and the SSRIs fluoxetine and sertraline have,
however, been found to bind in an extracellular vestibule termed
1
8,19
the S2 binding site of the bacterial protein LeuTAa
.
Data support-
ing the existence of a similar S2 site in hSERT have been pub-
3
20–23
mechanism.
lished,
but its role in transport and transport inhibition has
been debated. It has been shown that the S2 site might play an
allosteric role in the pharmacological action of serotonin itself.24
Binding to the S2 site can also affect the pharmacological properties
⇑
Corresponding authors.
E-mail addresses: steffen.sinning@clin.au.dk (S. Sinning), hhj@chem.au.dk
of some SSRIs and TCAs, in particular imipramine, where binding to
(
H.H. Jensen).
24
the S2 site can modulate the dissociation rate from hSERT.
A
y
These authors contributed equally.
Present address: D. E. Shaw Research, 120 W. 45th St., 39th Fl. New York, NY
detailed study has revealed that the S2 site might be located in an
extracellular vestibule approximately 10–12 Å above the S1
à
1
0036, United States.
http://dx.doi.org/10.1016/j.bmc.2016.04.039
968-0896/Ó 2016 Elsevier Ltd. All rights reserved.
0

2
726
A. Brinkø et al. / Bioorg. Med. Chem. 24 (2016) 2725–2738
binding site, and it has been suggested that the modulation of dis-
sociation rate via an allosteric site is due to a steric blockage of the
exit pathway caused by the low-affinity binding of antidepressants
on the several crystal structures of LeuTAa containing ligands
2
1,34–37
within the S2 site.
Additionally we demonstrate that it is
possible to position 3-substituted imipramine analogs that do
indeed fit into a hSERT homology model occupying both the S1
and the S2 site. The biological studies performed on wt hSERT
and selected mutants revealed that the shortest linkers yielded
the most potent inhibitors, and that they most likely are occupying
the S1 and S2 sites simultaneously.
2
0
to the S2 site.
The suggested existence of a S2 site in hSERT with affinity for
antidepressants indicates that it may be possible to synthesize
ligands that are dimeric or bifunctional in nature, and so could
occupy both S1 and S2 binding sites simultaneously. Literature
holds several examples of dimeric inhibitors, or inhibitors with
two interlinked binding motifs, showing improved affinity com-
pared to the single binder,25–30 and so a dimeric/bifunctional
antidepressant might possess improved binding affinity and selec-
tivity compared to other known antidepressant drugs. In a recent
study it was shown that citalopram analogs linked either symmet-
rically through the 5-position or un-symmetrically linked through
the N- and 5-position lost affinity compared to S-citalopram but
retained nanomolar affinity. Some of these dimeric compounds
were still able to bind to the S1 and to the allosteric site and even
exhibited improved allosteric properties, however, it remains
unclear whether their binding utilized both S1 and S2
2. Results
2.1. Steered MD
Similar to what has previously been explored for leucine (un)
3
6
binding in LeuTAa we performed Steered Molecular Dynamics
(SMD) simulations of the unbinding of imipramine from the central
S1 site of hSERT initiated from the previously biochemically vali-
9
dated binding mode. This methodology has proven useful in
exploring the relative stability of binding sites within a binding
event. Based on two 25 ns SMD simulations of imipramine from
the central S1 binding site of a dimer hSERT to the extracellular
space we were able to identify at least one secondary transient
binding site. In all four hSERT monomers we identified a secondary
binding site within the extracellular vestibule approximately 10–
12 Å above the central S1 site (Fig. 2). This site corresponds to
3
1
simultaneously.
In order to explore the possibility of targeting hSERT with a
ligand binding to both the S1 and S2 binding sites, we performed
computational studies which suggested that a well-defined bind-
ing site is located in an extracellular vestibule, approximately
1
0–12 Å from the S1 site. Accordingly, if an S2 site exists it will
likely be situated in this vestibule, consistent with previously pub-
the site previously described to bind several compounds within
9
,20,32
21,38–40
lished studies.
We have formerly biochemically validated a
LeuTAa
lations of LeuTAa
and which has also been identified from SMD simu-
3
6,38
computational study that placed imipramine in the primary S1
binding site orientated in such a way that the 3-position (Fig. 1)
was directed towards the exit of the binding pocket, in the direc-
tion of the extracellular vestibule9 and shown that extended sub-
stituents are tolerated at this position.37 This binding location and
.
From this secondary binding site we could
get an estimate of the binding distance between two imipramine
molecules (Fig. 2). Based on the predicted favorable binding sites
of imipramine we estimated an advantageous distance between
the two imipramine molecules to be ꢀ10 Å, and based on this we
designed novel molecular fragments that could possible occupy
both binding sites simultaneously, thereby possessing improved
binding affinity over a single imipramine alone. Previously, we also
observed that imipramine docked into the S2 site was orientated
with the alkyl amine pointing towards the central S1 binding site.9
,33
orientation has recently been confirmed by Gouaux and co-work-
ers using crystal structures of dDAT and a mutated LeuT.1
8,19
Given
the confidence we now have in our SERT model of imipramine
binding we set out to explore the existence and position of the
debated S2 site by synthesizing a series of ligands that could target
both binding sites, and experimentally challenge or support the
computational results. We decided to use iminodibenzyl fragments
2.2. Induced fit docking
to probe S2 because earlier studies have shown that tricyclic
antidepressants bind to the S2 site of LeuT.2
1,38
Furthermore,
Several antidepressants have been shown to also possess allos-
3
9,40
because the putative S2 site is expected to be smaller in hSERT
we also employed naphthyl fragments.
In this paper we report on both computational and experimen-
tal data supporting an S2 site to be located approximately 10–12 Å
from the S1 site, similar to what was observed from steered molec-
teric properties,
and it has been suggested that they exert their
20
effect by binding to a vestibular site located above the S1 site. We
hypothesized that a ligand utilizing both sites with a linker of the
appropriate length could have improved potency and selectivity.
As no details exist about the possible orientation of a antidepres-
sant analog in S2 we chose a linker with high flexibility. To test
our hypothesis we conceived a series of imipramine based ligands
consisting of an imipramine moiety attached via a linker to a naph-
thyl or iminodibenzyl fragment (Figs. 1 and 4). We chose the imin-
odibenzyl moiety because it is found in tricyclic antidepressants,
36
ular dynamics (SMD) of leucine unbinding in LeuTAa and based
Parent structure
9
which bind to the S1 in hSERT and may also have affinity for
1
0
11
9
1
the putative S2 site in hSERT based on the affinity of TCAs for S2
2
1
8
2
in LeuT. However, the putative S2 site in hSERT may also be con-
siderably smaller than in LeuT because LeuT L29, believed to line
the S2 site, corresponds to the much larger W103 in hSERT. To
compensate for the possibility of a smaller S2 site in hSERT we also
chose the aromatic naphthyl moiety instead of iminodibenzyl for a
series of the bifunctional inhibitors.
5
N
3
7
4
R
6
N
Induced fit docking calculations were performed for two of
these novel compounds, 8 and 12 (Fig. 4), to explore their binding
to hSERT. We chose these two compounds to isolate and explore
the effect of linker length in docking simulations. No poses
were obtained from the docking of compound 12, probably due
to the relatively long linker, which might accommodate greater
R = H, Imipramine, 1
R = Cl, Clomipramine, 2
R = -Linker-R', 6-21, this study
Figure 1. Structure of imipramine, clomipramine and the imipramine analogs 6–21
of this study.

A. Brinkø et al. / Bioorg. Med. Chem. 24 (2016) 2725–2738
2727
Figure 2. (A) Force profile (brown colors) and z-coordinate (green colors) during steered molecular dynamics as a function of time. Build-up in force indicates a favorable
interaction site. In all simulations the S1 site (z = 0) is the most favorable site and at least one second transient site is observed approximately 10 Å above the S1 site similar to
the previously described S2 site. (B) Position of imipramine (orange) in the initial S1 site and the proposed next stable S2 site obtained from the SMD calculations.
O
O
N
N
8
Figure 3. All four poses of compound 8 (green) within hSERT. Residues mutated in this study are shown in fat light brown sticks while selected residues are shown in gray
sticks. Only residue positions from the highest scoring pose are shown. The position of the binding site is shown on the left and zoom in on the binding site in two different
orientations are shown in the middle and to the right.
conformational changes within the protein. We were however able
to dock 8 and the binding mode of the imipramine moiety strongly
resembles the binding mode previously validated.9 Four poses
were obtained from the docking calculations and the binding with
the position of the naphthyl-group varying the most between
them. All four binding modes can be seen in Figure 3. The naph-
thyl-group of 8 is located in the S2 site similar to what was
observed from the steered MD simulations, placing this group in
close proximity to the salt bridge (Arg104-Glu493) within the
extracellular vestibule. The average GScore from the docking calcu-
lations is ꢁ15.5 kcal/mol. It is evident that the 3-linked molecules
are able to occupy both the S1 and S2 sites simultaneously in
induced fit docking calculations.
We speculated whether it would be possible to use the chlorine
atom of the easily obtainable pharmaceutical drug clomipramine
(2) as a chemical handle for attachment of the linker through a
Sonogashira coupling reaction with the linker functionalized with
a terminal alkyne. In literature, most examples of Sonogashira cou-
plings with chlorides utilize simple, activated aryl chlorides as
4
1,42
starting material.
In comparison, clomipramine (2) is a func-
tionalized aryl chloride, and accordingly only very few examples
exists of similarly demanding Sonogashira couplings.4
3,44
As a test
system the reaction was initially explored with clomipramine 2
and propargyl alcohol (Scheme 1) using conditions inspired by
4
5,46
work of the Buchwald group.
As no reaction took place, a sec-
ond substrate investigated was the tetrahydropyran (THP)-pro-
tected propargyl alcohol, and to our delight this gave the desired
Sonogashira coupled product in an acceptable yield of 63%. We
were unable to improve the yield by addition of a Cu(I) co-catalyst
or by altering either the base, temperature, or reaction medium.
Satisfied with this initial synthetic result we settled on a series
of ligands to be prepared to validate our computational results. The
structure of the linker and the R’ group fragments attached to imi-
pramine at the 3-position (Fig. 1) are shown in Figure 4. Varying
lengths of ethylene glycols or b-propylene glycol were chosen as
2
.3. Organic synthesis
Based on the promising results from the computational studies
we set out to synthesize a series of imipramine ligands to be
explored in biological test with hSERT and validate the computa-
9
,37
tional results. According to our imipramine binding model
it
was decided to attach the linker to the 3-position of one of the
benzo-moieties resulting in the general ligand design shown in
Figure 1, 6–21.
0
linkers to separate the imipramine fragment from the R -group,

2
728
A. Brinkø et al. / Bioorg. Med. Chem. 24 (2016) 2725–2738
Alkyne fragments:
O
O
O
S
O
O
O
6
7
8
9
O
O
O
O
O
O
O
O
O
S
1
0
11
12
13
O
N
O
O
N
1
4
15
Alkane fragments:
O
O
O
O
O
1
6
17
18
O
O
O
N
O
O
N
1
9
20
21
Figure 4. R group structure. Compound number refers to the whole structure, and not just the shown fragment.
OR'
NH
Cl
Pd
XPhos precat. PdG1 (3%)
Cs CO CH CN, 80 oC
P
N
N
Cl
2
3,
3
iPr
OR
N
N
iPr
iPr
XPhos precat. PdG1
O
O
O
O
,
R' = H, THP (3),
4
5
Scheme 1. Exploration of clomipramine (2) as a Sonogashira substrate.

A. Brinkø et al. / Bioorg. Med. Chem. 24 (2016) 2725–2738
2729
(
Fig. 4). As mentioned, ligands bearing both a naphthyl and the
protected propargyl alcohol is a privileged substrate in this reac-
tion. As a consequence of these failed reactions it was decided to
carry on with a linear synthesis by deprotection and extension of
the key intermediate 3. This required the synthesis of a series of
fragments that would carry a good leaving group and take part in
a Williamson ether synthesis with propargyl alcohol 22 (Scheme 2).
Accordingly, the THP-group was removed by treatment with
excess pTsOH in methanol which afforded alcohol 22 in 72% yield.
The ether forming reaction between this key intermediate and
either an iodide or tosylate functionalized linker would then result
in the first series of 3-functionalized alkyne linked imipramines
(Scheme 2). A detailed description of the linker synthesis and
attachment can be found in Section 5.
The reverse strategy for ether formation was also explored (not
shown) where a good leaving group in the form of a mesylate was
to be attached to the key intermediate 22, and then substituted by
a linker bearing an alcohol. This route, however, proved not to be
feasible for obtaining 6–15.
When all alkyne linked ligands 6–15 had been successfully syn-
thesized, they were to be reduced to the corresponding alkane
ligands by catalytic hydrogenation. This approach (Path A,
Scheme 2) was only successful for the alkyne ligands 6, 14 and
0
iminodibenzyl skeleton of imipramine in place of R were planned
to be synthesized to assess the tolerance of steric bulk in the S2
site.
The proposed ligands (6–15, Fig. 4) cover all linker lengths from
five to eleven atoms between the imipramine moiety and the
0
naphthyl/iminodibenzyl R -group. Ligands 8 and 9 allowed us to
test whether there would be a difference in binding between a 1-
and a 2-linked naphthyl fragment. The presence of an alkyne func-
tionality would also allow us to easily expand our collection of
ligands by simple alkyne reduction to the corresponding alkanes
1
6–21.
We also conceived the molecules 22 and 23 (Fig. 5) to see the
effect of the alkyne function itself. Simple substituents like 3-Cl
9
37
or 3-CN and 3,7-di-CH
increased inhibitory potency.
3
have been found to lead to 2-6 fold
For the synthesis of the proposed ligands (6 and 12, Fig. 4) a
convergent approach was investigated involving direct Sono-
gashira cross coupling on alkynes 4 and 5. Using the exact same
experimental procedure as for THP-protected propargyl alcohol
failed to give the desired products 6 and 12 suggesting that THP-
1
5 to give the corresponding alkane ligands in 27%, 85% and 93%
yield, respectively. The remaining alkyne ligands could not be
reduced by this protocol so their corresponding alkanes were pre-
pared through reduction of propargyl alcohol 22 to the saturated
analog 24, followed by linker attachment as previously described
in the alkyne series (Path B, Scheme 2). Alkynes 8, 10, 11, 14 and
N
N
OH
N
N
1
5 were successfully converted to alkane ligands 17–19 in 5%,
9
%, 18% and 20% yield over two steps, respectively.
The simple ligand 23 was prepared from clomipramine in a
Sonogashira reaction with (Triisopropylsilyl)acetylene. Deprotec-
2
2
23
Figure 5. Structure of short 3-substituted imipramine 22 and 23.
tion using TBAF afforded the analog 23 as shown in Scheme 3.
Path A
N
N
N
pTsOH
CH OH
NaH X
R''
3
OH 40 oC, DMF
X= Cl, I, OTs
OTHP
O
R''
7
2%
N
N
N
3
22
6-15 (12-76%)
H2, Pd(OH) /C
2
H2, Pd(OH) /C
2
Path B
EtOAc
EtOAc
NaH X
R''
N
N
OH 40 oC, DMF
X= Cl, I, OTs
O
R''
N
N
2
4
16-21 (5-93%)
0
00
00
Scheme 2. Synthesis of target alkynes and alkanes through propargyl alcohol. –Linker-R = CCCH
2
2
OCH R
or (CH
)
2 3
OCH
2
R .
Si
XPhos-precat (3%)
TBAF-THF
THF, rt
N
N
N
Cl
Cs CO CH CN, 85 o
C
2
3,
3
Si
N
N
N
25 (87%)
23 (83%)
Scheme 3. Synthesis of analog 23.

Table 1
3
i
Mean K values (nM) measured for inhibition of [ H]-5-HT uptake by HEK-293-MSR cells transiently transfected with hSERT wt or mutants
Compound
R-group structure
Wt hSERT (nM) K
31 [10;97], n = 6
i
(95%)
W103A (nM) IC50 (95%)
I179C (nM) K
i
(95%)
i
I172M (nM) K (95%)
1
AH
2.6 [1.6;4.3], n = 5
5.7 [2.1;15], n = 6
1300 [890;1900], n = 4
6
7
590 [290;1200], n = 4
1100 [630;1900], n = 4
340 [60;1800], n = 4
350 [140;860], n = 4
280 [74;1100], n = 3
5400 [1000;28,000], n = 3
O
S
O
475 [290;790], n = 5
4700 [2200;10,000], n = 4
O
8
9
3500 [770;16,000], n = 3
1500 [870;2600], n = 4
3500 [1500;8200], n = 4
4800 [4100;5600], n = 3
440 [110;1700], n = 3
940 [330;2700], n = 3
1800 [1000;3400], n = 3
640 [210;1900], n = 4
1200 [340;4200], n = 4
8000 [1800;34,000], n = 3
2000 [840;4600], n = 4
O
O
O
O
O
1
1
1
1
0
1
2
3
27,000 [14,000;53,000], n = 4
O
1100 [490;2600], n = 4
6500 [1700;25,000], n = 3
2000 [960;4300], n = 3
300 [230;390], n = 4
420 [180;1000], n = 3
560 [450;680], n = 4
540 [260;110], n = 3
450 [220;900], n = 3
500 [280;900], n = 4
5200 [3300;8400], n = 4
6400 [1300;31,000], n = 3
19,000 [7000;50,000], n = 3
O
O
S
O
O
O
O
1
4
13,000 [6800;24,000], n = 3
1600 [370;6600], n = 3
1600 [240;10,000], n = 3
16,000 [3800;65,000], n = 3
N
N
O
1
5
6
11,000 [6800;17,000], n = 3
620 [170;2300], n = 3
4100 [1400;12,000], n = 4
610 [280;1300], n = 4
2800 [1500;5200], n = 4
290 [120;670], n = 3
10,000 [4000;26,000], n = 4
5000 [1200;22,000], n = 3
O
O
1
O

O
O
17
18
19
600 [220;1700], n = 3
1200 [630;1900], n = 5
1000 [650;1700], n = 4
280 [90;860], n = 3
600 [190;1900], n = 5
280 [90;890], n = 3
270 [130;550], n = 3
620 [250;1500], n = 5
360 [250;500], n = 4
2700 [2100;3500], n = 4
9400 [4400;20,000], n = 5
6700 [2900;15,000], n = 5
O
O
O
O
2
0
1
6300 [2400;16,000], n = 3
5900 [3000;12,000], n = 3
270 [150;480], n = 3
840 [330;2300], n = 3
330 [90;1200], n = 3
640 [150;2700], n = 3
15000 [6100;68,000], n = 3
11000 [1500;86,000], n = 3
N
N
O
O
O
2
2
2
3
150 [50;420], n = 3
220 [110;460], n = 4
60 [20;170], n = 3
40 [8;160], n = 3
35 [24;52], n = 4
27 [12;63], n = 3
2600 [1400;4900], n = 5
2400 [1200;4900], n = 4
HO
2
M i
W103A was not active enough to yield a K , making conversion of W103A IC50 to K impossible. For this mutant we report an IC50. 95% confidence limits are shown in brackets.

2
732
A. Brinkø et al. / Bioorg. Med. Chem. 24 (2016) 2725–2738
With the series of target molecules in hand, covering different
chain lengths, and being a collection of 12 alkyne and 6 alkane
ligands, we set out to investigate their ability to inhibit hSERT.
N
N
2
.4. Inhibitory potencies of novel dimeric inhibitors for hSERT
O
S1 and S2 site mutants
N
N
n
The aim of the biochemical experiments was to evaluate the
inhibitory potency of the bi-functional ligands in a SAR-study by
uptake inhibition experiments on wt hSERT. Secondly, we wanted
to use selected mutants located in the S1 and S2 site to reveal
whether the ligands occupy the S1 and/or S2 site to qualify
whether a utilization of both sites had positive effects on potency
and would be a desirable strategy for inhibitor design. Inhibitory
potencies are shown in Table 1. As seen in Figure 3, W103 and
I179 are located in the putative S2 that would be occupied by
the 3-linked moiety of the compounds if they utilized both sites,
while I172 is located within the central S1 binding site. The loca-
tion of I172 in the central substrate site (later coined S1) was
established by Rudnick and co-workers many years prior to the
n=1, (IC50 6000 nM) 2, (IC50 4800 nM) 3 (IC50 5500 nM)
Figure 6. Inhibitory potency of nitrogen interlinked imipramine dimers.
taining ligands 6–13 and 16–19 (group mean = 1510 nM) when
compared as two groups (p <0.0001, two-tailed t-test). They inhibit
wt hSERT to the same extent as we previously observed for nitro-
5
0
gen interlinked imipramine dimers (Fig. 6).
Modification of the ligand may alter the preference for binding
sites, i.e., where imipramine prefers the S1 site, a bivalent imipra-
mine analog may now prefer both the S1 and S2 sites or it may only
prefer S2. We have included three mutants in the study to enable
such classification, one in S1 and two in the larger and more versa-
tile S2. The effect of these mutations on inhibitor potency relative
to hSERT wt will aid us in determining which sites are utilized by
the ligands.
4
7
first structural insights. Blakely and co-workers have demon-
strated a key role of this residue in coordinating high affinity
4
8
inhibitors and proceeded to establish an antidepressant-insensi-
49
tive knock-in hSERT I172M mouse. By determining the orienta-
tion of an SSRI and a TCA in the S1 site9 we have previously
11
Trp103 is a sterically demanding residue located within the
proposed S2 site. When mutating this residue to the much smaller
alanine, we do however also see a positive effect on the potency of
almost all the 3-coupled ligands. The larger ligands do obtain a
relative larger gain in affinity and for example ligand 14 obtains
an 8-fold increase in affinity (p = 0.004, two-tailed t-test), while
molecule 20 obtains a 23-fold increase in affinity (p = 0.0002,
two-tailed t-test). The only molecule out of the 20 compounds
not showing a large increase in potency upon the W103A mutation
is compound 8, which could possibly be caused by favorable
aromatic interactions between the naphthyl moiety and W103 as
seen for the docking poses (Fig. 3).
been able to show how I172 was nested into the curved heteroaro-
matic substructure of imipramine, readily explaining the detri-
mental effects of the I172M mutation on imipramine potency
and suggesting how the sensitivity to the I172M mutation could
be a marker of S1 site utilization by a ligand. Large bi-functional
inhibitors, as in the current study, might in some instances not
reach the S1 and therefore we used the sensitivity to I172M as a
marker for S1 site binding.
2
.5. Inhibitory potencies of 3-linked imipramine analogs for
inhibition of the human serotonin transporter
The residue Ile179 is considered to be located in the proposed
S2 site, thus mutation of this to the smaller Cys will increase the
size of the S2 binding pocket and likely also the funnel connecting
S1 and S2. Interestingly, the ligands containing the iminodibenzyl
fragments (14, 15, 20 and 21) were affected positively by the
I179C mutation, becoming significantly more potent (hSERT wt
group mean = 8500 nM vs hSERT I179C group mean = 1060 nM,
p = 0.0084).
In an initial SAR analysis we were interested in whether the
introduced substitutions affected analog potency relative to the
parent compound, imipramine. Overall, we found that all the stud-
ied analogs exhibited significantly decreased potency for inhibition
of hSERT wt relative to imipramine (p <0.05, one-way ANOVA,
Dunnett’s post hoc test).
The ligands 22 and 23 (Table 1) with small 3-substituents show
that the presence of an alkyne function itself at this position signif-
icantly decreases binding to wt hSERT as opposed to other small
substituents like 3-Cl and 3-CN which have earlier been found to
Ile172 is positioned within the central S1 binding site and
mutation of this residue into methionine decreased the potency
for almost all molecules (hSERT wt group mean = 2100 nM vs
hSERT I179C group mean = 7100 nM, p <0.0001). The fact that
almost all ligands are affected by the S1 single point mutation indi-
cates that almost all molecules are occupying the central binding
pocket forming favorable interactions with Ile172 (Fig. 3), which
9
increase potency. Increasing the size of the 3-substituent by going
to bi-functional ligands with the shortest alkyne linker results in
further lowering of inhibitory potency, i.e., the potency of 6 is sig-
nificantly decreased relative to 23. The slightly longer alkane linker
of 16 yields a very similar potency as 6 and is also significantly
decreased relative to 24, which is the most relevant reference com-
pound. These were, together with alkane 17, however the most
potent of the bi-functional ligands. The naphthyl displaying ligands
having longer linkers are generally worse inhibitors of wt hSERT.
The optimal point of attachment of the linker to the naphthyl
group is most directly assessed by the comparison of the 2-naphthyl
displaying ligand 9 and the 1-naphthyl displaying ligand 8, which
have identical linker length and linker chemistry. Although the
5
2
accordingly are destroyed by the I172M mutation. The loss of
affinity can potentially be due to steric clashes introduced by the
slightly longer methionine side chain than seen in wt and/or loss
of hydrophobic interactions between Ile and imipramine within
the wt protein. However, a group of structurally diverse bivalent
ligands (compounds 9, 12, 14, 15) show little or no loss of affinity
in the I172M mutant relative to wt hSERT which could be a sign
that these ligands do either not utilize the S1 site for binding or
that they do so in a different orientation than imipramine.
2
-substituted 9 appears to have higher potency than 8, this falls
short of statistical significance.
We utilized two different classes of substituents at the end of
the linker. Viewed as a group, the larger iminodibenzyl containing
ligands 14, 15, 20, 21 (group mean = 8500 nM), were found to be
significantly less potent against wt hSERT than the naphthyl con-
3. Conclusions
The computer simulations clearly illustrate the presence of a
transient S2 site in hSERT at a position previously described within

A. Brinkø et al. / Bioorg. Med. Chem. 24 (2016) 2725–2738
2733
LeuT21,36,38–40,42 however, with a lower affinity than observed for
the central binding site as judged from the SMD simulations. Based
on the spatial organization between the S1 and S2 site we were
able to construct 3-linked imipramine analogs which were addi-
tionally tested through molecular docking, clearly illustrating that
these bi-functional ligands are able to occupy the central and
extracellular binding sites simultaneously. Based on these results
the synthesis of 3-linked imipramine ligands to be used for biolog-
ical studies of hSERT, was undertaken. In total 19 ligands were syn-
thesized using a copper-free Sonogashira reaction as the key step
and the drug clomipramine as the starting material. This is the first
example of the Sonogashira reaction being used to directly modify
a known drug, providing easy access to analogs hereof. The phar-
macological data shows that the shortest alkyne ligand 6 was the
most potent bi-functional ligand against wt hSERT. When the
ligands were tested against W103A and I179C mutants, enlarging
the S2 binding site, it was found that the larger ligands were
affected positively, and to a larger extent than the smaller ligands,
suggesting that the ligands are occupying the proposed S2 site, as
also predicted from the computational studies. Additionally, the
I172M mutation within the central binding pocket negatively
affected almost all of the tested molecules which indicated that
almost all the novel molecules are indeed occupying both the cen-
tral S1 and the extracellular S2 site simultaneously.
dimer structure of LeuTAa.6 The dimer was manually embedded
into a pre-equilibrated POPC membrane bilayer. The system was
solvated and neutralized to a 0.2 M ion concentration using NaCl.
4.3. Simulations details
The simulations were performed utilizing the NAMD pro-
5
5,56
57
gram
applying the CHARMM22 force field for protein and
5
8
lipids including the CMAP corrections. The same parameters for
imipramine as previously described by Sinning et al. were used.
9
5
9
Particle Mesh Ewald was used for long range electrostatic and
the cutoff for the van der Waals interactions was 12 Å with a
switching function applied from 10 Å. Constant pressure at 1 atm
was maintained utilizing the Langevin piston method60 with a
decay time of 50 fs and an oscillation period of 100 fs. The temper-
ature was held constant at 310 K by Langevin damping with a
ꢁ1
damping coefficient of 0.5 f s
.
In pulling simulations, a moving constraint is applied to the
center of mass (COM) of a group of atoms. The chosen group of
atoms will thus be forced to generally move along a specified direc-
tion while still being able to move freely along all other degrees of
freedom. Constant velocity SMD was accomplished by the tclforces
6
0,61
in NAMD.
The ligand SMD simulations were performed as pre-
viously described with a pulling force of 500 pN/Å2 and with a
62
The S2 site could be envisaged to have both positive and nega-
tive effects on ligand affinity depending on the ligand. If the S2 site
serves as a transient vestibular binding site en route to S1 it would
increase the local concentration of ligand experienced by the S1
site and increase association rates which in turn would increase
potency. If the S2 site contributes negatively by impeding access
to S1 then we would expect that enlarging the S2 site would lead
to increased imipramine potency. We do indeed observe that
expanding the S2 site with the W103A and I179C mutations does
increase imipramine potency suggesting that the S2 site con-
tributes negatively to imipramine potency.
pulling velocity of 1 Å/ns.
4
.4. Imipramine docking
Ligand 8 and 12 (Fig. 4) were docked into the combined S1 and
S2 binding sites of the protein utilizing induced fit docking proto-
63
64
col (IFD ) which utilizes Glide from the Schrödinger suite as pre-
9,12–14,16,17,58
viously done by us.
In short, the binding site was
defined around Ile172 and Asp98. The number of poses to save
was set to 100 in both initial and re-docking stage. The XP-scoring
65
function was used in the redocking stage.
Taken together, our findings serve as the starting point for
design and development of a new generation of improved antide-
pressants that fully exploit both binding sites, and hence, possibly
can achieve a better inhibitory potency than the mono-functional
counterpart and possibly achieve an inhibitor that is non-compet-
itive with 5-HT through additional use of the S2. The level of inhi-
bition obtained with the ligands described in this article could
possibly be greatly improved by exchanging the naphthyl fragment
with another moiety that contributes more to binding.
4
.5. Mutagenesis
The cDNA for the human serotonin transporter (Uniprot
P31645) in the pCDNA 3.1 vector was used as template for the
mutagenesis. Mutations were introduced by PCR using the Phusion
High-Fidelity DNA polymerase (Finnzymes) and primer pairs with
appropriate nucleotide mismatches followed by DpnI digestion of
the parent DNA. Escherichia coli XL10 (Stratagene) were trans-
formed with the mutated DNA and used for DNA production.
Mutant constructs were sequenced across the entire reading frame
using BigDye v3.1 chemistry (Applied Biosystems) analyzed on an
ABI 3100 Sequencer (Applied Biosystems) to ensure that the trans-
porter gene contained the desired mutations and that no unwanted
mutations had been introduced.
4
4
. Methods
.1. Homology modeling
The hSERT model utilized in the study is the one previously
1
1,15,17
published.
In summary, this model was constructed utilizing
5
1,52
6
MODELLER 9v4
based on the structure of LeuTAa utilizing the
4.6. Cell culture
alignment by Beuming et al.53 including extensive optimization of
EL2 along with two sodium ions and one chloride ion placed as
Human embryonic kidney 293MSR cells (Invitrogen) were cul-
tured in Dulbecco’s modified Eagle’s medium (BioWhitaker) sup-
plemented with 10% fetal calf serum (Invitrogen), 100 U/ml
5
4
described previously.
4
.2. Simulation system building
penicillin, 100
lg/mL streptomycin (BioWhitaker), and 6
lg/mL
Geneticin (Invitrogen) at 95% humidity and 5% p(CO
2
) at 37 °C.
The structure used for the steered molecular dynamics simula-
Two days prior to the uptake assay, cells were detached by Versene
(Invitrogen) and trypsin/EDTA (BioWhitaker) treatment and mixed
with a preformed complex of 0.2 lg transporter DNA and 0.5 lL
Lipofectamine 2000 per square centimeter of plating area
(Invitrogen). The transfection mix was dispensed into white
96-well microplates (Nalgene Nunc International) at a cell density
of 50–70% confluence and incubated for 48–60 h.
tions was hSERT with imipramine bound in the same orientation as
9
biochemically validated and later confirmed from crystal struc-
1
8,19
tures.
The steered molecular dynamics simulations were set
up in a similar way as for the SMD simulations of the (un)binding
events of leucine from LeuTAa (PDB 2A65) by Celik et al.3 A dimer
of hSERT with imipramine bound was constructed based on the
6

2
734
A. Brinkø et al. / Bioorg. Med. Chem. 24 (2016) 2725–2738
4
.7. Uptake inhibition assay
3 4
saturated NaHCO (aq). The organic phase was dried over MgSO ,
filtered and evaporated. The product was purified by column
chromatography.
Adherent transfected cells were washed with phosphate-buf-
fered saline supplemented with calcium and magnesium (PBSCM:
1
37 mM NaCl, 2.7 mM KCl, 10 mM Na
2
HPO
4
*H
2
O, 1.4 mM KH
2
PO
4
,
5.1.3. General procedure C: Coupling procedure to the
formation ligands 17–19
0
.1 mM CaCl , and 1 mM MgCl , pH 7.4) and immediately preincu-
2
2
bated for 30 min with a dilution series of the drug. Initiation of
uptake was performed by adding a mix of 50–100 nM [ H] 5-HT
The alcohol 22 (1 equiv) was dissolved in anhydrous DMF to a
concentration of 1.9 mol/L under a N atm. before NaH (60%,
2
3
and the inhibitor in the same concentration as in the preincuba-
tion. This incubation lasted for 10 min at room temperature and
was terminated by aspiration and washing with PBSCM. The
amount of accumulated radioactive neurotransmitter was deter-
mined by lysing the cells with Microscint 20 (PerkinElmer) and
quantified on a Packard Topcounter.
2 equiv) and the appropriate iodide (1.5 equiv) were added. The
reaction mixture was heated to 40 °C. When the reaction had run
to completion (app. 3 h), it was quenched with H
with saturated NaHCO (aq). The aqueous phase was extracted
with AcOEt and washed alternately with H O and saturated
NaHCO (aq). The organic phase was dried over MgSO , filtered
2
O and diluted
3
2
3
4
and evaporated. The product was purified by column
chromatography.
4
.8. Data analysis
Analysis of all data was carried out in GraphPad Prism 5.0 soft-
ware (GraphPad Inc., San Diego, CA). Radioactive counts from [ H]
5.1.4. General procedure D: Catalytic hydrogenation to the
formation of compounds 16, 20, 21, 24
3
5
-HT were fitted by sigmoidal dose–response curve for determina-
The appropriate alkyne was dissolved in AcOEt to a concentra-
tion of IC50-values. The IC50-values were converted to K
i
-values
tion of 0.03 mol/L, under N
added. The flask was evacuated and backfilled with H
left overnight. When crude H NMR showed complete reduction of
the alkyne to the alkane, the mixture was filtered on a Celite col-
2
atm, before Pd(OH)
2
/C (20 mol %) was
(g), and then
using the Cheng–Prusoff equation.66
2
1
One-way ANOVA with Dunnett’s post hoc test and the Student
t-test was used for statistical comparison of K values.
Ò
i
umn and the solvent was evaporated. In most cases there was no
need for further purification.
5
5
. Experimental
.1. General methods
5.1.5. N,N-Dimethyl-3-(3-(3-(tetrahydro-2H-pyran-2-yl)oxy)
prop-1-yn-1-yl)-10,11-dihydro-5H-dibenzo[b,f]azepine-5-yl)
propane-1-amine, 3
Moisture sensitive reactions were carried out in oven (ca.
20 °C) or flame dried glassware, under a N or Ar atm, in solvents
1
2
Clomipramine 2 (0.503 g; 1.43 mmol; 1 equiv) was mixed with
Cs CO (1.13 g; 3.57 mmol; 2.5 equiv) and XPhos precatalyst PdG1
dried according to standard procedures. TLC analysis was per-
2
3
formed on aluminum sheets coated with silica gel (Merck kiesel,
(0.0313 g; 0.0428 mmol; 3 mol %) in a glove box. To this anhydrous
acetonitrile (4 mL) was added, and the reaction mixture was left
with stirring, in the glove box for 25 min before THP protected
propargyl alcohol (0.501 g; 3.57 mol; 2.5 equiv) was added. The
reaction vessel was sealed, removed from the glove box, and then
heated to 80 °C for 20 h. After 20 h the reaction mixture was cooled
6
0, F254) and were visualized by either UV-irradiation or staining
with a KMnO (KMnO and NaOH in H O) or Cmol ((NH Mo
O, Ce(SO and H SO in H O) solution, and heated until spots
4
4
2
4
)
6
7 24
O -
ꢂ4H
2
4
)
2
2
4
2
appeared. Flash column chromatography was performed using
1
Merck silica 60 (230–400 mesh) as stationary phase. H NMR, and
1
3
C NMR were recorded on a Varian Mercury 400 spectrometer,
to rt and quenched with saturated NaHCO (aq). The mixture was
3
at 400 MHz, and 100 MHz respectively. Mass spectral data were
recorded on a Micromass LC-TOF instrument as electrospray
experiments. Melting points were measured on a Büchi B-450,
and are not corrected.
extracted with AcOEt, and the combined organic phases were dried
over MgSO , filtered and evaporated. The product was purified by
4
column chromatography (first CH Cl /CH OH 50:1 then CH Cl /
2
2
3
2
2
CH OH 20:1). Upon repeated column chromatography, the product
3
was isolated as brown oil in 63% yield (0.375 g; 0.895 mmol). R
(CH Cl /CH OH 20:1) 0.45. H NMR (400 MHz, CDCl ) dH 7.16–
2 2 3 3
f
1
5
.1.1. General procedure A: Coupling procedure to the
formation of ligands 6, 8, 12–15
7.02 (m, 4H), 6.99 (s, 2H), 6.94 (dt, 1H, J = 7.3, 1.1 Hz), 4.88 (t,
1H, J = 3.4 Hz) 4.46 (q, 2H, J = 15.7 Hz), 3.93–3.81 (m, 1H), 3.77 (t,
2H, J = 6.6 Hz), 3.61–3.50 (m, 1H), 3.12 (s, 4H), 2.60 (t, 2H,
The alcohol 22 (1 equiv) was dissolved in anhydrous DMF to a
concentration of 0.37 mol/L, under a N
equiv) and the appropriate iodide or tosylate in case of 6
1.3 equiv) were added. The reaction mixture was heated to 40 °C
and left overnight. When the reaction had run to completion, it
was quenched with H O and diluted with saturated NaHCO (aq).
The aqueous phase was extracted with ethyl acetate and washed
alternately with H O and saturated NaHCO (aq). The organic
phase was dried over MgSO , filtered and evaporated. The product
was purified by column chromatography.
2
atm, before NaH (60%,
1
3
2
(
J = 5.6 Hz), 2.35 (s, 6H), 1.96–1.48 (m, 8H). C NMR (100 MHz,
CDCl ) d 147.9, 134.9, 134.4, 130.3, 129.7, 126.7, 125.9, 123.3,
3
C
123.2, 120.7, 120.4, 96.9, 86.0, 84.4, 62.1, 57.3, 54.9, 48.5, 44.9,
32.6, 31.8, 30.4, 25.5, 19.2 ppm. HRMS (ES+): Calcd for C27H34N O -
2
3
2
2
H: 419.2699; found 419.2699.
2
3
4
5.1.6. N,N-Dimethyl-3-(3-(3-(naphthalene-1-ylmethoxy)prop-1-
yn-1-yl)-10,11-dihydro-5H dibenzo[b,f]azepine-5-yl)propane-1-
amine, 6
5
.1.2. General procedure B: Coupling procedure to the
The product 6 was synthesized according to general procedure
A, and was purified by column chromatography (first CH Cl /CH -
formation of ligands 9–11
2
2
3
The alcohol 22 (1 equiv) was dissolved in anhydrous DMF to a
2 2 3
OH 50:1 then CH Cl /CH OH 10:1). The product was obtained as
concentration of 1.9 mol/L, under a N
2
atm, before NaH (60%,
f 2 2 3
yellow oil in 31% yield (0.029 g; 0.0619 mmol). R (CH Cl /CH OH
10:1) 0.54. H NMR (400 MHz, CDCl ) d 8.23 (d, 1H, J = 8.1 Hz),
3 H
1
2
equiv) and the appropriate iodide (3 equiv) were added to the
solution. The reaction mixture was heated to 40 °C. When the reac-
tion had run to completion (approx. 3 h), it was quenched with
7.86 (dd, 2H. J = 15.4, 8.1 Hz), 7.61–7.39 (m, 4H), 7.22 (s, 1H),
7.18–7.06 (m, 3H), 7.04 (s, 2H), 6.96 (dd, 1H, J = 16.1, 9.0 Hz),
5.14 (s, 2H), 4.45 (s, 2H), 3.77 (t, 2H, J = 6.9 Hz), 3.16 (s, 4H), 2.32
H
2
O and diluted with saturated NaHCO
3
(aq). The aqueous phase
O and
(t, 2H, J = 7.2 Hz), 2.16 (s, 6H), 1.73 (quint, 2H, J = 6.8 Hz). 13
C
was extracted with AcOEt and washed alternately with H
2

A. Brinkø et al. / Bioorg. Med. Chem. 24 (2016) 2725–2738
2735
NMR (100 MHz, CDCl
3
) d
30.3, 129.6, 129.0, 128.6, 127.3, 126.7, 126.5, 125.9, 125.6, 125.3,
24.3, 123.3, 123.1, 120.6, 120.5, 87.0, 84.3, 70.1, 58.1, 57.7, 48.9,
C
148.2, 134.9, 134.6, 133.9, 133.1, 132.0,
1.89 (t, 2H, J = 7.2 Hz). 13C NMR (100 MHz, CDCl
147.6, 134.5, 130.4, 129.8, 129.5, 127.7, 126.9, 126.8, 126.5,
126.0, 123.8, 123.5, 123.0, 120.7, 120.4, 119.1, 106.9, 86.7, 84.3,
68.4, 67.4, 59.5, 57.0, 48.2, 44.3, 32.6, 31.7, 24.3. HRMS (ES+): Calcd
3
) d
C
156.8, 147.7,
1
1
4
4
5.6, 32.7, 31.8, 26.2 ppm. HRMS (ES+): Calcd for C33
75.2749; found 475.2791.
34 2
H N OH:
37 2 2
for C34H N O H: 505.2855; found 505.2853.
5
.1.7. N,N-Dimethyl-3-(3-(3-((naphthalene-1-ylthio)methoxy)
5.1.10. N,N-Dimethyl-3-(3-(3-(3-(naphthalene-1-yloxy)propoxy)
prop-1-yn-1-yl)-10,11-dihydro-5H-dibenzo[b,f]azepine-5-yl)
propane-1-amine, 10
prop-1-yn-1-yl)-10,11-dihydro-5H-dibenzo[b,f]azepine-5-yl)
propane-1-amine, 7
DBU (0.570 g; 3.70 mmol; 1.2 equiv) was added to a solution of
The product 10 was synthesized according to general procedure
1
-thionaphthalene (0.500 g; 3.12 mmol; 1 equiv) dissolved in
2 2 3
B, and purified by column chromatography (CH Cl /CH OH 50:1).
anhydrous CH Cl (27 mL) under a N atmosphere, and left over-
night. When the reaction had run to completion the organic phase
was washed with 1 M HCl and brine. The product (chloromethyl)
2
2
2
The product was obtained as yellow oil in 12% yield (0.012 g;
1
f 2 2 3
0.0226 mmol). R (CH Cl /CH OH 10:1) 0.50. H NMR (400 MHz,
CDCl ) d 8.27 (d, 1H, J = 8.0 Hz), 7.78 (d, 1H, J = 7.8 Hz), 7.51–
3
H
(
naphthalene-1-yl)sulfane (0.032 g; 0.151 mmol; 1 equiv), still dis-
solved in CH Cl , was then added to a solution of alcohol 22
0.051 g; 0.151 mmol; 1 equiv) dissolved in anhydrous DMF
0.2 mL). CH Cl was then evaporated and NaI (0.023 g;
.151 mmol; 1 equiv) and NaH (0.012 g; 0.303 mmol; 2 equiv)
were added to the reaction mixture. After 3 h the reaction had
run to completion, and was quenched carefully with H O. The
aqueous phase was diluted with saturated NaHCO (aq) and
extracted with AcOEt. The combined organic phases were dried
over MgSO , filtered and evaporated. The product was purified by
column chromatography (CH Cl /CH OH 50:1.5) and isolated as
yellow oil in 27% yield (0.021 g; 0.0407 mmol). R (CH Cl /CH OH
8.42 (d, 1H, J = 8.4 Hz),
.85 (dd, 2H, J = 7.3, 4.7 Hz), 7.78 (d, 1H, J = 8.2 Hz), 7.62–7.47
7.31 (m, 4H), 7.18–7.01 (m, 4H), 7.01–6.89 (m, 3H), 6.84 (d, 1H,
J = 7.4 Hz), 4.40 (s, 2H), 4.29 (t, 2H, J = 6.1 Hz), 3.88 (t, 2H,
J = 6.2 Hz), 3.71 (t, 2H, J = 6.8 Hz), 3.14 (s, 4H), 2.37–2.21 (m, 4H),
2
2
(
(
0
1
3
2
2
2.16 (s, 6H), 1.71 (quint, 2H, J = 7.2 Hz). C NMR (100 MHz, CDCl
3
)
d
C
154.8, 148.1, 134.9, 134.6, 134.5, 130.2, 129.6, 127.5, 126.6,
126.4, 126.0, 125.7, 125.2, 123.2, 123.1, 122.1, 120.5, 120.2,
104.8, 86.6, 84.3, 67.0, 65.1, 59.2, 57.7, 48.8, 45.5, 32.6, 31.8,
2
3
29.8, 26.0. HRMS (ES+): Calcd for C35
519.3015.
38 2 2
H N O H: 519.3012; found
4
2
2
3
5.1.11. N,N-Dimethyl-3-(3-(3-(2-(2-(naphthalene-1-yl)ethoxy)
ethoxy)prop-1-yn-1-yl)-10,11-dihydro-5H-dibenzo[b,f]azepine-
5-yl)propane-1-amine, 11
f
2
2
3
1
1
7
3 H
0:1) 0.48. H NMR (400 MHz, CDCl ) d
The product 11 was synthesized according to general procedure
(
m, 2H), 7.44 (t, 1H, J = 8.0 Hz), 7.18–7.05 (m, 4H), 7.05–6.89 (m,
2 2 3
B, and purified by column chromatography (CH Cl /CH OH 50:1).
3
2
H), 5.26 (s, 2H), 4.63 (s, 2H), 3.75 (t, 2H, J = 6.9 Hz), 3.15 (s, 4H),
.33 (t, 2H, J = 7.2 Hz), 2.18 (s, 6H), 1.73 (quint, 2H, J = 7.2 Hz).
The product was obtained as yellow oil in 28% yield (0.027 g;
1
0.0501 mmol). R
f 2 2 3
(CH Cl /CH OH 50:4) 0.55. H NMR (400 MHz,
1
3
C NMR (100 MHz, CDCl
) d
3 C
148.1, 134.9, 134.7, 134.1, 133.3,
CDCl ) d 8.07 (d, 1H, J = 8.3 Hz), 7.84 (d, 1H, J = 9.3 Hz), 7.75–
3
H
1
1
5
C
32.7, 130.3, 129.8, 129.6, 128.7, 128.1, 126.7, 126.6, 126.3,
25.9, 125.8, 125.2, 123.3, 123.2, 120.5, 120.3, 87.2, 83.2, 74.4,
7.6, 55.9, 48.8, 45.4, 32.7, 31.8, 25.9 ppm. HRMS (ES+): Calcd for
7.69 (m, 1H), 7.53–7.42 (m, 2H), 7.42–7.34 (m, 2H), 7.20–7.08
(m, 3H), 7.08–6.94 (m, 4H), 4.42 (s, 2H), 3.87–3.67 (m, 8H), 3.41
(t, 2H, J = 7.6 Hz), 3.12 (s, 4H), 2.86 (dd, 2H, J = 9.2, 6.7 Hz), 2.49
1
3
33
H
34
N
2
OSH: 507.2407; found: 507.2468.
(s, 6H), 2.14–1.94 (m, 2H). C NMR (100 MHz, CDCl
3 C
) d 147.4,
1
47.2, 134.8, 134.3, 133.9, 132.2, 130.6, 129.9, 128.9, 127.1,
5
.1.8. N,N-Dimethyl-3-(3-(3-((naphthalene-1-yloxy)ethoxy)
127.0, 126.9, 126.3, 126.1, 125.7, 125.6, 123.8, 122.8, 120.9,
120.3, 86.3, 84.8, 71.8, 70.3, 69.4, 59.4, 56.5, 47.7, 43.4, 33.4,
prop-1-yn-1-yl)-10,11-dihydro-5H-dibenzo[b,f]azepine-5-yl)
propane-1-amine, 8
32.5, 31.7 22.9. HRMS (ES+): Calcd for C36
found 533.3165.
40 2 2
H N O H: 553.3168;
The product 8 was synthesized according to general procedure
A, and was purified by column chromatography (first CH
OH 50:1 then CH Cl /CH OH 10:1). The product was obtained as
reddish/brown oil in 24% yield (0.034 g; 0.0709 mmol). R (CH Cl
8.33 (dd, 1H,
J = 8.2, 1.5 Hz), 7.79 (dd, 1H, J = 7.4, 1.4 Hz), 7.52–7.32 (m, 4H),
2 2 3
Cl /CH -
2
2
3
5.1.12. N,N-Dimethyl-3-(3-(3-(2-(2-(naphthalene-1-yloxy)
ethoxy)ethoxy)prop-1-yn-1-yl)-10,11-dihydro-5H-dibenzo[b,f]
azepine-5-yl)propane-1-amine, 12
f
2
2
/
1
3 3 H
CH OH 10:1) 0.53. H NMR (400 MHz, CDCl ) d
The product 12 was synthesized according to general procedure
7
2
.22–6.90 (m, 7H), 6.84 (d, 1H, J = 7.5 Hz), 4.56 (s, 2H), 4.37 (dd,
H, J = 5.5, 4.2 Hz), 4.12 (dd, 2H, J = 5.2, 4.5 Hz), 3.74 (t, 2H,
A, and purified by column chromatography (first CH
50:1 then CH Cl /CH OH 10:1). The product was obtained as yel-
low oil in 41% yield (0.068 g; 0.123 mmol). R (CH Cl /CH OH
8.34–8.26 (m, 1H),
2 2 3
Cl /CH OH
2
2
3
J = 6.9 Hz), 3.15 (s, 4H), 2.29 (t, 2H, J = 6.8 Hz), 2.15 (s, 6H), 1.71
quint, 2H, J = 7.2 Hz). 13C NMR (100 MHz, CDCl
) d 154.7, 148.2,
34.6, 130.3, 129.6, 127.5, 126.6, 126.5, 125.9, 125.7, 125.3,
f
2
2
3
1
(
3
C
3 H
10:1) 0.68. H NMR (400 MHz, CDCl ) d
1
1
5
C
7.85–7.75 (m, 1H), 7.52–7.31 (m, 4H), 7.19–6.90 (m, 7H), 6.85–
6.78 (m, 1H) 4.45 (s, 2H), 4.34 (t, 2H, J = 5.2 Hz), 4.04 (t, 2H,
J = 5.4 Hz), 3.89–3.86 (m, 2H) 3.83–3.80 (m, 2H), 3.73 (t, 2H,
J = 6.9 Hz), 3.14 (s, 4H), 2.30 (t, 2H, J = 6.8 Hz), 2.15 (s, 6H), 1.70
23.3, 123.1, 122.3, 120.6, 120.5, 105.0, 87.0, 84.1, 68.4, 67.9,
9.6, 57.7, 48.9, 45.6, 32.7, 31.8, 26.2. HRMS (ES+): Calcd for
34
H
36
N
2
O
2
H: 505.2855; found 505.2853.
1
3
(
3 C
quint, 2H, J = 6.8 Hz). C NMR (100 MHz, CDCl ) d 154.6, 147.8,
5
.1.9. N,N-Dimethyl-3-(3-(3-(3-(2-(naphthalene-2-yloxy)ethoxy)
134.8, 134.4, 130.3, 129.7, 127.5, 126.7, 126.5, 125.9, 125.3,
123.3, 123.1, 122.2, 120.7, 120.5, 120.4,105.1, 86.6, 84.4, 70.9,
69.9, 69.3, 68.0, 59.4, 57.2, 48.4, 44.7, 32.6, 31.7, 24.8. HRMS (ES
prop-1-yn-1-yl)-10,11-dihydro-5H-dibenzo[b,f]azepine-5-yl)
propane-1-amine, 9
The product 9 was synthesized according to general procedure
40 2 3
+): Calcd for C36H N O H: 549.3117; found 549.3126.
B, and was purified by column chromatography (CH
2 2 3
Cl /CH OH
5
0:1.5). The product was isolated as yellow oil in 28% yield
5.1.13. N,N-Dimethyl-3-(3-(3-(2-(2-(naphthalene-1-ylthio)
ethoxy)ethoxy)prop-1-yn-1-yl)-10,11-dihydro-5H-dibenzo[b,f]
azepine-5-yl)propane-1-amine, 13
1
(
0.026 g; 0.0517 mmol). R
f 2 2 3
(CH Cl /CH OH 10:1) 0.43. H NMR
(
400 MHz, CDCl ) d 7.74 (dd, 3H, J = 17.6, 8.4 Hz), 7.43 (t, 1H,
3
H
J = 7.5 Hz), 7.33 (t, 1H, J = 7.5 Hz), 7.22–7.08 (m, 6H), 7.08–6.92
The product 13 was synthesized according to general procedure
(
(
m, 3H), 4.51 (s, 2H), 4.34–4.27 (m, 2H), 4.10–3.98 (m, 2H), 3.75
t, 2H, J = 6.6 Hz), 3.13 (s, 4H), 2.62 (t, 2H, J = 7.6 Hz), 2.34 (s, 6H),
A, and purified by column chromatography (first CH
2 2 3
Cl /CH OH
50:1 then CH Cl /CH OH 10:1). The product was obtained as
2
2
3

2
736
A. Brinkø et al. / Bioorg. Med. Chem. 24 (2016) 2725–2738
orange/brown oil in 45% yield (0.048 g; 0.0841 mmol). R
CH ) d 8.45 (d, 1H,
OH 10:1) 0.46. 1H NMR (400 MHz, CDCl
J = 8.3 Hz), 7.84 (d, 1H, J = 9.1 Hz), 7.74 (d, 1H, J = 8.2 Hz), 7.65 (d,
H, J = 7.2 Hz), 7.53 (dddd, 2H, J = 20.5, 8.0, 6.8, 1.3 Hz), 7.44–7.37
f
(CH
2
Cl
2
/
5.1.17. N,N-Dimethyl-3-(3-(3-(2-(naphthalene-1-yloxy)ethoxy)
propyl)-10,11-dihydro-5H-dibenzo[b,f]azepine-5-yl)propane-1-
amine, 17
3
3
H
1
The product 17 was synthesized according to general procedure
(
(
m, 1H), 7.19–6.91 (m, 7H), 4.40 (s, 2H), 3.77–3.62 (m, 8H), 3.20
C, and purified by column chromatography (CH
2 2 3
Cl /CH OH 50:1).
t, 2H, J = 7.0 Hz), 3.14 (s, 4H), 2.31 (t, 2H, J = 7.2 Hz), 2.16 (s, 6H),
The product was obtained as yellow oil in 10% yield (9 mg;
.71 (quint, 2H, J = 6.0 Hz). 13C NMR (100 MHz, CDCl
) d
148.2,
0.0169 mmol). R
1
1
1
1
8
2
5
3
C
f 2 2 3
(CH Cl /CH OH 50:4) 0.51. H NMR (400 MHz,
34.9, 134.6, 134.1, 133.3, 133.1, 130.2, 129.6, 129.0, 128.7,
27.7, 126.6, 126.4, 125.7, 125.6, 125.3, 123.3, 123.1, 120.5, 86.8,
4.1, 70.4, 70.1, 69.2, 59.4, 57.7, 48.9, 45.6, 33.9, 32.6, 31.8,
CDCl ) d 8.31 (d, 1H, J = 9.6 Hz), 7.79 (d, 1H, J = 7.7 Hz), 7.52–
3
H
7.31 (m, 3H), 7.12–6.80 (m, 8H), 6.74 (d, 1H, J = 9.1 Hz), 4.31 (t,
2H, J = 5.2 Hz), 3.94 (t, 2H, J = 5.2 Hz), 3.73 (t, 2H, J = 6.9 Hz), 3.62
(t, 2H, J = 6.4 Hz), 3.12 (s, 4H), 2.66 (t, 2H, J = 7.6 Hz), 2.30 (t, 2H,
J = 7.2 Hz), 2.14 (s, 6H), 1.92 (t, 2H, J = 6.4 Hz), 1.70 (quint, 2H,
6.2. HRMS (ES+): Calcd for C36
65.2903.
40 2 2
H N O SH: 565.2889; found
1
3
J = 7.2 Hz). C NMR (100 MHz, CDCl
3
) d
34.7, 129.9, 129.8, 127.5, 126.5, 126.0, 125.3, 122.6, 122.3,
20.6, 120.2, 120.1, 105.1, 70.9, 69.5, 68.1, 57.8, 48.9, 45.6, 32.4,
C
154.8, 148.5, 140.2,
1
1
3
5
5
.1.14. 3-(3-(3-(3-(10,11-Dihydro-5H-dibenzo[b,f]azepine-5-yl)
propoxy)prop-1-yn-1-yl)-10,11-dihydro-5H-dibenzo[b,f]
azepine-5-yl)-N,N-dimethylpropane-1-amine, 14
2.2, 32.1, 31.6, 26.2. HRMS (ES+): Calcd for C34
09.3168; found 509.3167.
40 2 2
H N O H:
The product 14 was synthesized according to general procedure
A, and purified by column chromatography (first CH
0:1 then CH Cl /CH OH 10:1). The product was obtained as yel-
low oil in 76% yield (0.120 g; 0.211 mmol). R (CH Cl /CH OH
7.19–7.04 (m, 9H),
.04–6.86 (m, 6H), 4.30 (s, 2H), 3.87 (t, 2H, J = 6.5 Hz), 3.75 (t,
H, J = 6.6 Hz), 3.62 (t, 2H, J = 6.7 Hz), 3.14 (s, 8H), 2.31 (t, 2H,
2 2 3
Cl /CH OH
5
.1.18. N,N-Dimethyl-3-(3-(3-(3-(naphthalene-1-yloxy)propoxy)
5
2
2
3
propyl)-10,11-dihydro-5H-dibenzo[b,f]azepine-5-yl)propane-1-
amine, 18
f
2
2
3
1
1
7
2
3 H
0:1) 0.28. H NMR (400 MHz, CDCl ) d
The product 18 was synthesized according to general procedure
2 2 3
C, and purified by column chromatography (CH Cl /CH OH 50:1.5).
The product was obtained as light yellow oil in 10% yield (0.014 g;
0.0272 mmol). R (CH Cl /CH OH 10:1) 0.51. H NMR (400 MHz,
J = 6.8 Hz), 2.16 (s, 6H), 1.90 (quint, 2H, J = 6.3 Hz), 1.72 (quint,
1
1
3
1
1
1
4
C
H, J = 6.8 Hz). C NMR (100 MHz, CDCl
) d
3 C
148.4, 148.2, 134.9,
f
2
2
3
CDCl
.31 (m, 4H), 7.15–6.98 (m, 3H), 6.97–6.81 (m, 4H), 6.70 (d, 1H,
J = 7.6 Hz), 4.27 (t, 2H, J = 6.1 Hz), 3.72 (q, 4H, J = 7.2 Hz), 3.47 (t,
H, J = 6.4 Hz), 3.11 (s, 4H), 2.61 (t, 2H, J = 8.0 Hz), 2.32 (t, 2H,
J = 7.2 Hz), 2.27–2.11 (m, 8H), 1.88 (quint, 2H, J = 7.2 Hz), 1.72
3 H
) d 8.29 (d, 1H, J = 9.0 Hz), 7.80 (d, 1H, J = 7.2 Hz), 7.52–
34.5, 134.4, 130.2, 129.9, 129.6, 126.6, 126.5, 125.7, 123.3,
23.0, 122.5, 120.6, 120.5, 120.1, 86.5, 84.4, 68.1, 59.0, 57.7, 48.9,
7.6, 45.7, 32.6, 32.3, 31.8, 28.3, 26.3. HRMS (ES+): Calcd for
7
2
39
H
43
N
3
OH: 570.3484; found 570.3508.
1
3
(
1
1
quint, 2H, J = 7.6 Hz). C NMR (100 MHz, CDCl
3
) d
48.3, 140.2, 134.6, 131.4, 129.9, 129.8, 127.6, 126.5, 126.1,
25.2, 122.6, 122.2, 120.2, 120.1, 104.8, 70.4, 67.6, 65.2 (, 57.8
C
154.9, 148.4,
5
.1.15. 3-(3-(3-(3-(3-(10,11-Dihydro-5H-dibenzo[b,f]azepine-5-
yl)propoxy)propoxy)prop-1-yn-1-yl)-10,11-dihydro-5H-dibenzo
[b,f]azepine-5-yl)-N,N-dimethylpropane-1-amine, 15
(
1C, 16), 48.9, 45.5, 32.4, 32.3, 32.0, 31.6, 30.0, 26.1. HRMS (ES+):
The product 15 was synthesized according to general procedure
Calcd for C35 H: 523.3325; found 523.3325.
42 2 2
H N O
A, and was purified by column chromatography (first CH
OH 50:1 then CH Cl /CH OH 10:1). The product was obtained as
light yellow oil in 40% yield (0.067 g; 0.107 mmol). R (CH Cl /CH
7.19–7.05 (m, 10H),
.02–6.87 (m, 5H), 4.26 (s, 2H), 3.82 (t, 2H, J = 6.8 Hz), 3.74 (t,
2 2 3
Cl /CH
2
2
3
5
.1.19. N,N-Dimethyl-3-(3-(3-(2-(2-(naphthalene-1-yl)ethoxy)
f
2
2
3
ethoxy)propyl)-10,11-dihydro-5H-dibenzo[b,f]azepine-5-yl)
propane-1-amine, 19
1
3 H
OH 10:1) 0.48. H NMR (400 MHz, CDCl ) d
7
2
3
1
1
1
6
The product 19 was synthesized according to general procedure
H, J = 6.9 Hz), 3.58 (t, 2H, J = 6.4 Hz), 3.47 (quint, 4H, J = 6.3 Hz),
.16 (s, 8H), 2.31 (t, 2H, J = 7.2 Hz), 2.16 (s, 6H), 1.84 (m, 4H),
C, and purified by column chromatography (CH
The product was obtained as light yellow oil in 10% yield (0.012 g;
2 2 3
Cl /CH OH 50:1.5).
.71 (quint, 2H, J = 6.4 Hz). 13C NMR (100 MHz, CDCl
) d
3 C
148.4,
1
0
.0216 mmol). R
f f 2 2 3
R (CH Cl /CH OH 10:1) 0.50. H NMR (400 MHz,
48.2, 134.9, 134.5, 134.3, 130.2, 129.9, 129.6, 126.6,126.5, 125.6,
23.3, 123.1, 122.5, 120.6, 120.5, 120.1, 86.3, 84.5, 68.7, 67.8,
7.3, 59.0, 57.7, 48.9, 47.6, 45.6, 32.6, 32.4, 31.8, 30.1, 28.4, 26.2.
CDCl ) d) d 8.08 (d, 1H, J = 8.3 Hz), 7.85 (d, 1H, J = 7.7 Hz), 7.73 (dd,
3
H
1
7
H, J = 6.9, 2.3 Hz), 7.49 (dt, 2H, J = 13.4, 6.7 Hz), 7.43–7.35 (m, 2H),
.14–7.02 (m, 3H), 6.99 (d, 1H, J = 7.7 Hz), 6.95–6.86 (m, 2H), 6.75
HRMS (ES+): Calcd for C42
49 3 2
H N O H: 628.3903; found 628.3958.
(
d, 1H, J = 7.6 Hz), 3.84 (t, 2H, J = 7.6 Hz), 3.76 (t, 2H, J = 6.8 Hz),
3
3
6
.69–3.55 (m, 4H), 3.47 (t, 2H, J = 6.5 Hz), 3.41 (t, 2H, J = 7.6 Hz),
5
.1.16. N,N-Dimethyl-3-(3-(3-(naphthalene-1-ylmethoxy)
.12 (s, 4H), 2.62 (t, 2H, J = 7.6 Hz), 2.38 (t, 2H, J = 7.6 Hz), 2.19 (s,
H), 1.88 (quint, 2H, J = 7.2 Hz), 1.76 (quint, 2H, J = 7.2 Hz). 13
C
propyl)-10,11-dihydro-5H-dibenzo[b,f]azepine-5-yl)propane-1-
amine, 16
3 C
NMR (100 MHz, CDCl ) d 148.3, 140.3, 134.9, 134.7, 134.0, 132.3,
The product 16 was hydrogenated following general procedure
1
1
3
31.4, 129.9, 128.9, 127.1, 126.9, 126.5, 126.0, 125.7, 125.6,
23.9, 122.6, 120.2, 120.0, 71.8, 70.9, 70.5, 70.3, 57.7, 48.8, 45.3,
3.5, 32.4, 32.3, 32.1, 31.4, 25.9. HRMS (ES+): Calcd for C36H N O
44 2 2
D, and purified by column chromatography (first CH
0:1 then CH Cl /CH OH 10:1). The product was obtained as color-
less oil in 27% yield (9 mg; 0.0186 mmol). R (CH Cl /CH OH 15:1)
8.14 (d, 1H, J = 8.1 Hz), 7.85 (dd,
2 2 3
Cl /CH OH
5
2
2
3
f
2
2
3
H: 537.3481; found 537.3481
1
0
2
1
3 H
.47. H NMR (400 MHz, CDCl ) d
H, J = 23.7, 8.8 Hz), 7.57–7.40 (m, 4H), 7.13–7.02 (m, 3H), 6.97 (d,
H, J = 7.7 Hz), 6.93–6.87 (m, 2H), 6.69 (dd, 1H, J = 7.7, 1.6 Hz), 4.96
5
.1.20. 3-(3-(3-(3-(10,11-Dihydro-5H-dibenzo[b,f]azepine-5-yl)
propoxy)propyl)-10,11-dihydro-5H-dibenzo[b,f]azepine-5-yl)-
N,N-dimethylpropane-1-amine, 20
The product was synthesized according to general procedure D,
and isolated as yellow oil in 85% yield (0.062 g; 0.108 mmol). R
(
2
s, 2H), 3.72 (t, 2H, J = 6.9 Hz), 3.58 (t, 2H, J = 6.3 Hz), 3.13 (s, 4H),
.64 (t, 2H, J = 7.4 Hz), 2.28 (t, 2H, J = 7.0 Hz), 2.13 (s, 6H), 1.93
13
(
(
1
1
3
quint, 2H, J = 6.4 Hz), 1.69 (quint, 2H, J = 6.0 Hz).
100 MHz, CDCl ) d 148.5, 148.3, 140.2, 134.7, 134.2, 133.9,
31.9, 131.4, 129.9, 129.8, 128.6, 126.4, 126.2, 125.9, 125.4,
24.2, 122.5, 120.2, 120.1, 71.6, 69.9, 57.9, 48.9, 45.7, 32.4, 32.3,
2.1, 31.7, 26.3. HRMS (ES+): Calcd for C33 OH: 479.3062;
C NMR
f
1
3
C
(
2 2 3 3 H
CH Cl /CH OH 15:1) 0.32. H NMR (400 MHz, CDCl ) d 7.16–
7
3
.05 (m, 10H), 6.99–6.86 (m, 4H), 6.68 (dd, 1H, J = 7.6, 1.6 Hz),
.85 (t, 2H, J = 6.8 Hz), 3.75 (t, 2H, J = 6.9 Hz), 3.45 (t, 2H,
38 2
H N
J = 6.2 Hz), 3.37 (t, 2H, J = 6.4 Hz), 3.21–3.08 (m, 8H), 2.55 (t, 2H,
J = 7.6 Hz), 2.32 (t, 2H, J = 7.2 Hz), 2.16 (s, 6H), 1.90–1.76 (m, 4H),
found 479.3115.

A. Brinkø et al. / Bioorg. Med. Chem. 24 (2016) 2725–2738
2737
1
1
1
3
5
.76–1.67 (m, 2H). 13C NMR (100 MHz, CDCl
40.3, 134.6, 134.3, 131.4, 129.9, 129.8, 126.5, 126.4, 122.5,
20.2, 120.1, 120.0,70.3, 68.7, 57.8, 48.9, 47.8, 45.5, 32.4, 32.3,
2.0, 31.5, 28.5, 26.1. HRMS (ES+): Calcd for
74.3797; found 574.3867.
3
) d
C
148.4, 148.3,
5.1.24. 3-(3-ethynyl-10,11-dihydro-5H-dibenzo[b,f]azepin-5-yl)-
N,N-dimethylpropan-1-amine 23
TBAF in THF (1.0 M, 0.15 mL) was added dropwise over a period
of 5 min to a solution of 25 (76.3 mg, 0.166 mmol, 1 eq) in THF
39 47 3
C H N OH:
(1.65 mL) at 0 °C under a N
for 24 h before the solution was diluted with dichloromethane
50 mL), washed with 10% NaHCO , dried over MgSO , filtered
2
atmosphere. The reaction was stirred
(
3
4
5
.1.21. 3-(3-(3-(3-(3-(10,11-Dihydro-5H-dibenzo[b,f]azepine-5-
and evaporated. The crude product was purified by flash column
chromatography (ethyl acetate followed by 2.5% methanol in
dichloromethane followed by 5.0% methanol in dichloromethane
yl)propoxy)propoxy)propyl)-10,11-dihydro-5H-dibenzo[b,f]
azepine-5-yl)-N,N-dimethylpropane-1-amine, 21
The product was synthesized according to general procedure D,
and was isolated as light yellow oil in 93% yield (0.031 g;
followed by 10.0% methanol in dichloromethane) to yield 24 as
1
OH 15:1) 0.48. 1H NMR (400 MHz,
brown oil (60.5 mg, 83%) R (Methanol/Ethyl acetate 1:1) 0.23.
f
H
0
.0491 mmol). R
CDCl ) d 7.17–7.05 (m, 9H), 7.00 (d, 1H, J = 7.7 Hz), 6.94–6.88
m, 4H), 6.74 (dt, 1H, J = 9.2, 4.6 Hz), 3.84 (t, 2H, J = 6.8 Hz), 3.77
t, 2H, J = 6.9 Hz), 3.51–3.39 (m, 6H), 3.37 (t, 2H, J = 6.4 Hz), 3.20–
.07 (m, 8H), 2.60 (t, 2H, J = 6.4 Hz), 2.31 (t, 2H, J = 6.4 Hz), 2.16
f
2
(CH Cl
2
/CH
3
3
NMR (400 MHz, CDCl ): d 7.25–6.90 (m, 7H) 3.77 (t, 2H, J = 6 Hz)
3
H
3
.14 (s, 4H) 3.01 (s, 1H) 2.42 (t, 2H, J = 8 Hz) 2.23 (s, 6H) 1.79
(
(
3
1
3
(
quint, 2H, J = 8 Hz). C NMR (100 MHz, CDCl
33.7, 129.2–118.9, 82.8, 75.2, 56.3, 47.5, 44.0, 31.5, 30.6, 24.3.
HRMS (ES+): Calcd for C21H24N H: 305.2018 m/z, found m/z:
3
): d 146.9 146.8,
1
1
3
2
(
s, 6H), 1.83 (dquint, 6H, J = 12.8, 6.4 Hz), 1.77–1.67 (m, 2H).
C
3
05.2045.
3 C
NMR (100 MHz, CDCl ) d 148.4, 140.3, 134.7, 134.3, 131.4,
1
5
29.9, 129.8, 126.5, 126.4, 122.5, 120.1, 70.3, 68.7, 68.0, 67.9,
7.9, 49.0, 47.7, 45.7, 32.4, 32.3, 32.1, 31.6, 30.3, 28.4, 26.4. HRMS
H: 628.4216 found 628.4220
5.1.25. 3-(5-(3-(Dimethylamino)propyl)-10,11-dihydro-5H-
dibenzo[b,f]azepine-3-yl)propane-1-ol, 24
(
ES+): Calcd for C42
H
49
N
3
O
2
The product 24 was hydrogenated following general procedure
D, and was isolated as yellow oil in 93% yield (0.087 g;
5
.1.22. 3-(5-(3-(Dimethylamino)propyl)-10,11-dihydro-5H-
1
0
.256 mmol). R
CDCl ) d
m, 2H), 6.76 (d, 1H, J = 7.6 Hz), 3.78 (t, 2H, J = 7.0 Hz), 3.59 (t,
f
(CH
2
Cl
2
/CH
3
OH 10:1) 0.31. H NMR (400 MHz,
dibenzo[b,f]azepine-3-yl)prop-2-yn-1-ol 22
H
7.17–7.05 (m, 3H), 7.01 (d, 1H, J = 7.7 Hz), 6.97–6.86
p-Toluenesulfonic acid (0.296 g; 1.56 mmol; 1.9 equiv) was
added to a solution of 3 (0.343 g; 0.820 mmol; 1 equiv) dissolved
in methanol (8.5 mL). After 2 h the reaction had run to completion
and was quenched with saturated NaHCO3(aq) before methanol was
3
(
2
H, J = 6.4 Hz), 3.15 (s, 4H), 2.64 (t, 2H, J = 7.6 Hz), 2.34 (t, 2H,
J = 7.2 Hz), 2.17 (s, 6H), 1.85 (quint, 2H, J = 8.0 Hz), 1.75 (quint,
1
3
2
1
4
3 C
H, J = 7.2 Hz). C NMR (100 MHz, CDCl ) d 148.6, 147.9, 140.1,
2
evaporated. The residue was diluted with H O and the aqueous
34.6, 131.3, 129.9, 129.7, 126.4, 122.5, 122.4, 120.1, 61.7, 57.7,
9.0, 45.5, 34.3, 32.3, 32.0, 31.9, 26.2. HRMS (ES+): Calcd for
phase was extracted with dichloromethane. The combined organic
phases were dried over magnesium sulfate, filtered and evapo-
rated. The product was purified by column chromatography (first
dichloromethane/methanol 50:1 then dichloromethane/methanol
22 30 2
C H N OH: 339.2436; found 339.2429.
Acknowledgement
1
0:1) and isolated as brown oil in 72% yield (0.196 g; 0.587 mmol).
1
R
CDCl
(
f
(dichloromethane/methanol 10:1) 0.29. H NMR (400 MHz,
) d 7.19–7.02 (m, 4H), 7.02–6.89 (m, 3H), 4.41 (s, 2H), 3.72
t, J = 6.8 Hz, 2H), 3.45 (s, 1H), 3.13 (s, 4H), 2.33 (t, J = 7.4 Hz, 2H),
B.S. and H.K. gratefully thank The Danish Council for Indepen-
dent Research Natural Sciences and the Lundbeck, Carlsberg
Foundations and the Alfred Benzon Foundation for financial
support along with the Danish Center for Scientific Computing
for access to supercomputing facilities. S.S. gratefully thanks the
Lundbeck Foundation for financial support.
3
H
1
3
2
(
1
3
3
.16 (s, 6H), 1.79–1.68 (qv, J = 6.8 Hz, 2H) ppm.
100 MHz, CDCl ) d 148.1, 148.0, 134.9, 134.2, 130., 129.5, 126.6,
25.4, 123.1, 123.0, 120.8, 120.4, 87.4, 85.1, 57.5, 51.0, 48.8, 45.2,
2.6, 31.7, 25.8 ppm. HRMS (ES+): Calcd for C22H N OH: m/z
26 2
35.2123; found m/z 335.2160.
C NMR
3
C
Supplementary data
5
.1.23. N,N-Dimethyl-3-(3-((triisopropylsilyl)ethynyl)-10,11-
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmc.2016.04.039.
dihydro-5H-dibenzo[b,f]azepine-5-yl)propane-1-amine, 25
In a glove box with argon atmosphere clomipramine (100.0 mg,
0
0
.285 mmol, 1 equiv), (triisopropylsilyl)acetylene (104.0 mg,
.569 mmol, 2 equiv), Cs CO (0.231 g, 0.712 mmol, 2.5 equiv),
References and notes
2
3
acetonitrile (1 mL) and XPhos-precatalyst PdG1 (6.31 mg,
.0085 mmol, 3%) were all mixed in a vial. Subsequently the vial
was closed, removed from the glove box and the reaction mixture
was stirred in a preheated heating block at 85 °C. After 7 h the
reaction was allowed to cool to room temperature, diluted with
1
.
Andersen, J.; Kristensen, A. S.; Bang-Andersen, B.; Strømgaard, K. Chem.
Commun. 2009, 3677.
0
2. Kristensen, A. S.; Andersen, J.; Jørgensen, T. N.; Eriksen, J.; Løland, C. J.;
Strømgaard, K. Pharmacol. Rev. 2011, 63, 585.
3
4
.
.
Wong, D. T.; Perry, K. W.; Bymaster, F. P. Nat. Rev. Drug Disc. 2005, 4, 764.
Stahl, S. M. Essential Psychopharmacology: the Prescriber’s Guide: Antipsychotics
and Mood Stabilizers; Cambridge University Press: New York, 2006.
Torres, G. E.; Gainetdinov, R. R.; Caron, M. G. Nat. Neurosci. 2003, 4, 13.
Yamashita, A.; Singh, S. K.; Kawate, T.; Jin, Y.; Gouaux, E. Nature 2005, 437, 215.
Singh, S. K. Channels 2008, 2, 380.
NaHCO
3
10% (10 mL) and extracted with ethyl acetate. The com-
, filtered and evapo-
5
6
7
.
.
.
bined organic phases were dried over MgSO
4
rated. The crude product was purified by flash column
chromatography (Dichloromethane followed by 5% methanol in
8. Wang, C.-I. A.; Lewis, R. J. Biochem. Pharmacol. 2010, 79, 1083.
9
.
Sinning, S.; Musgaard, M.; Jensen, M.; Severinsen, K.; Celik, L.; Koldsøe, H.;
Meyer, T.; Bols, M.; Jensen, H. H.; Schiøtt, B.; Wiborg, O. J. Biol. Chem. 2010, 285,
dichloromethane) to yield compound 26 as brown oil (128.1 mg,
1
8
7%).
R
f
(5% methanol in dichloromethane) 0.41.
H NMR
8363.
(
3
400 MHz, CDCl
3
): d 7.20–6.87 (m, 7H) 3.77 (t, 2H, J = 8.0 Hz)
.12 (s, 4H) 2.48 (t, 2H, J = 7.4 Hz) 2.27 (d, 6H) 1.82 (quint. 2H,
10. Andersen, J.; Taboureau, O.; Hansen, K. B.; Olsen, L.; Egebjerg, J.; Strømgaard,
C.; Kristensen, A. S. J. Biol. Chem. 2009, 284, 10276.
11. Koldsø, H.; Severinsen, K.; Tran, T. T.; Celik, L.; Jensen, H. H.; Wiborg, O.;
Schiøtt, B.; Sinning, S. J. Am. Chem. Soc. 2010, 132, 1311.
12. Severinsen, K.; Koldsø, H.; Thorup, K. A. V.; Schjøth-Eskensen, C.; Møller, P. T.;
Wiborg, O.; Jensen, H. H.; Sinning, S.; Schiøtt, B. Mol. Pharmacol. 2014, 85, 208.
13
J = 8 Hz) 1.13 (s, 21H, (triisopropylsilyl)).
CDCl
4
C NMR (100 MHz,
3
): d 148.5 148.3, 135.2, 135.0, 130.6–120.8, 107.9 90.1, 57.8,
9.1, 45.4, 33.0, 32.4, 25.7, 19.4, 12.0. MS

2
738
A. Brinkø et al. / Bioorg. Med. Chem. 24 (2016) 2725–2738
1
3. Andersen, J.; Stuhr-Hansen, N.; Zachariassen, L. G.; Koldsø, H.; Schiøtt, B.;
Strømgaard, K.; Kristensen, A. S. Mol. Pharmacol. 2014, 85, 703.
4. Severinsen, K.; Kraft, J. F.; Koldsø, H.; Vinberg, K. A.; Rothman, R. B.; Partilla, J.
S.; Wiborg, O.; Blough, B.; Schiøtt, B.; Sinning, S. ACS Chem. Neurosci. 2012, 3,
42. Littke, A. F.; Fu, G. C. Angew. Chem., Int. Ed. 2002, 41, 4176.
43. Merkul, E.; Oeser, T.; Muller, T. J. J. Chem. Eur. J. 2009, 15, 5006.
44. Lee, D. H.; Lee, Y. H.; Harrowfield, J. M.; Lee, I. M.; Lee, H. I.; Lim, W. T.; Kim, Y.;
Jin, M. J. Tetrahedron 2009, 65, 1630.
1
6
93.
45. Gelman, D.; Buchwald, S. L. Angew. Chem., Int. Ed. 2003, 42, 5993.
46. XPhos pre-catalyst (PdG1) was used instead of using separate phosphine and
palladium source, see Ref. 42.
47. Chen, J. G.; Rudnick, G. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 1044.
48. Henry, L. K.; Field, J. R.; Adkins, E. M.; Parnas, M. L.; Vaughan, R. A.; Zou, M.-F.;
Newman, A. H.; Blakely, R. D. J. Biol. Chem. 2006, 281, 2012.
1
1
5. Koldsø, H.; Autzen, H. E.; Grouleff, J.; Schiøtt, B. PLoS One 2013, 8, e63635.
6. Celik, L.; Sinning, S.; Severinsen, K.; Hansen, C. G.; Møller, M. S.; Bols, M.;
Wiborg, O.; Schiøtt, B. J. Am. Chem. Soc. 2008, 130, 2542.
1
7. Koldsø, H.; Christiansen, A. B.; Sinning, S.; Schiøtt, B. ACS Chem. Neurosci. 2013,
4
, 295.
1
8. Wang, H.; Goehring, A.; Wang, K. H.; Penmatsa, A.; Ressler, R.; Gouaux, E.
Nature 2013, 503, 141.
9. Penmatsa, A.; Wang, K. H.; Gouaux, E. Nature 2013, 503, 85.
0. Plenge, P.; Shi, L.; Beuming, T.; Te, J.; Newman, A. H.; Weinstein, H.; Gether, U.;
Loland, C. J. J. Biol. Chem. 2012, 287, 39316.
49. Thompson, B. J.; Jessen, T.; Henry, L. K.; Field, J. R.; Gamble, K. L.; Gresch, P. J.;
Carneiro, A. M.; Horton, R. E.; Chisnell, P. J.; Belova, Y.; McMahon, D. G.; Daws,
L. C.; Blakely, R. D. Proc. Natl. Acad. Sci. U.S.A. 2011, 108, 3785.
50. Bright, S. A.; Brinkø, A.; Larsen, M. T.; Sinning, S.; Williams, D. C.; Jensen, H. H.
Bioorg. Med. Chem. Lett. 2013, 23, 1220.
1
2
2
2
2
1. Singh, S. K.; Yamashita, A.; Gouaux, E. Nature 2007, 448, 952.
2. Plenge, P.; Gether, U.; Rasmussen, S. G. Eur. J. Pharmacol. 2007, 567, 1.
3. Loland, C. J.; Plenge, P.; Shi, L.; Zhang, P.; Javitch, J. A.; Newman, A. H.;
Weinstein, H.; Gether, U. Basic Clin. Pharmacol. Toxicol. 2010, 107, 116.
4. Plenge, P.; Mellerup, E. T.; Laursen, H. Eur. J. Pharm. 1991, 206, 243.
5. Susheck, S. J.; Wong, A. L.; Koeller, K. M.; Boehr, D. D.; Draker, K.; Sears, P.;
Wright, G. D.; Wong J. Am. Chem. Soc. 2000, 122, 5230.
51. Sali, A.; Blundell, T. L. Mol. Biol. 1993, 234, 779.
52. Fiser, A.; Do, R. K.; Scaronali, A. Protein Sci. 2000, 9, 1753.
53. Beuming, T.; Shi, L.; Javitch, J. A.; Wein, N. H. Mol. Pharmacol. 2006, 70, 1630.
54. Koldsø, H.; Severinsen, K.; Tran, T. T.; Celik, L.; Jensen, H. H.; Wiborg, O.;
Schiott, B.; Sinning, S. J. Am. Chem. Soc. 2010, 132, 1311.
2
2
55. Kalé, L.; Skeel, R.; Bhandarkar, M.; Brunner, R.; Gursoy, A.; Krawetz, N.; Phillips,
J.; Shinozaki, A.; Varadarajan, K.; Schulten, K. J. Comput. Phys. 1999, 151, 283.
56. Philips, J. C.; Braun, R.; Wang, W.; Gumbart, J.; Tajkhorshid, E.; Villa, E.; Chipot,
C.; Skeel, R. D.; Kalé, L.; Schulten, K. J. Comput. Chem. 2005, 26, 1781.
57. MacKerell, A. D. J.; Bashford, D.; Bellott, M.; Dunbrack, R. L., Jr.; Evanseck, J. D.;
Field, M. J.; Fischer, S.; Gao, J.; Guo, H.; Ha, S.; Joseph-McCarthy, D.; Kuchnir, L.;
Kuczera, K.; Lau, F. T. K.; Mattos, C.; Michnick, S.; Ngo, T.; Nguyen, D. T.;
Prodhorn, B.; Reiher, W. E., III; Roux, B.; Schlenkrich, M.; Smith, J. C.; Stote, R.;
Straub, J.; Watanabe, M.; Wiorkiewicz-Kuczera, J.; Yin, D.; Karplus, M. J. Phys.
Chem. B 1998, 102, 3586.
2
6. Bach, A.; Chi, C. N.; Pang, G. F.; Olsen, L.; Kristensen, A. S.; Jemth, P.;
Strømgaard, K. Angew. Chem., Int. Ed. 2009, 48, 9685.
2
7. Chan, K.-F.; Zhao, Y.; Burkitt, B. A.; Wong, I. L. K.; Chow, L. M. C.; Chan, T. H. J.
Med. Chem. 2006, 49, 6742.
2
2
3
8. Kramer, R. H.; Karpen, J. W. Nature 1998, 398, 710.
9. Tian, L.; Heyduk, T. Biochemistry 2009, 48, 264.
0. Schmitt, K. C.; Mamidyala, S.; Biswas, S.; Dutta, A. K.; Reith, M. E. A. J.
Neurochem. 2010, 112, 1605.
3
1. Banala, A. K.; Zhang, P.; Plenge, P.; Cyriac, G.; Kopajtic, T.; Katz, J. L.; Loland, C.
J.; Newman, A. H. J. Med. Chem. 2013, 56, 9709.
58. MacKerell, A. D.; Feig, M.; Brooks, C. L., III J. Comput. Chem. 2004, 24, 1400.
59. Darden, T.; York, D.; Pedersen, L.; Ewald, Particle mesh J. Chem. Phys. 1993, 98,
10089.
60. Feller, S. E.; Zhang, Y.; Pastor, R. W.; Brooks, B. R. J. Chem. Phys. 1995, 103, 4613.
61. Kalé, L.; Skel, R.; Bhandakar, M.; Brunner, R.; Gursoy, A.; Krawetz, N.; Phillips, J.;
Shinozaki, A.; Varadarajan, k.; Schulten, K. J. Comput. Phys. 1999, 151, 283.
62. Celik, L.; Schiott, B.; Tajkhorshid, E. Biophys. J. 2008, 94, 1600.
63. Sherman, W.; Day, T.; Jacobson, M. P.; Friesner, R. A.; Farid, R. J. Med. Chem.
2006, 49, 534.
3
3
2. Celik, L.; Schiøtt, B.; Tajkhorshid, E. Biophys. J. 2008, 94, 1600.
3. Christensen, H.; Schjøth-Eskesen, C.; Jensen, M.; Sinning, S.; Jensen, H. H. Chem.
Eur. J. 2011, 17, 10618.
3
4. Zhou, Z.; Zhen, J.; Karpowich, N. K.; Goetz, R. M.; Law, C. J.; Reith, M. E. A.;
Wang, D.-N. Science 2007, 317, 1390.
3
5. Zhou, Z.; Zhen, J.; Karpowich, N. K.; Law, C. J.; Reith, M. E. A.; Wang, D.-N. Nat.
Struct. Mol. Biol. 2009, 16, 652.
3
3
3
3
4
6. Singh, S.; Piscitelli, C. L.; Yamashita, A.; Gouaux, E. Science 2008, 322, 1655.
7. Quick, M.; Winther, A. L.; Shi, L.; Nissen, P.; Weinstein, H. PNAS 2009, 106, 5563.
8. Shi, L.; Quick, M.; Zhao, Y.; Weinstein, H.; Javitch, J. A. Mol. Cell 2008, 30, 667.
9. Plenge, P.; Mellerup, E. Eur. J. Pharmacol. 1985, 119, 1.
64. Suite 2011: Schrödinger Suite 2011 Induced Fit Docking protocol; Glide
Version 5.7, Schrödinger, LLC, New York, NY, 2011; Prime Version 2.2,
Schrödinger, LLC, New York, NY, 2011.
65. Friesner, R. A.; Murphy, R. B.; Repasky, M. P.; Frye, L. L.; Greenwood, J. R.;
Halgren, T. A.; Sanschagrin, P. C.; Mainz, D. T. J. Med. Chem. 2006, 49, 6177–
6196.
0. Biessen, E. A.; Norder, J. A.; Horn, A. S.; Robillard, G. T. Biochem. Pharmacol.
1
988, 37, 3959.
4
1. Torborg, C.; Huang, J.; Schulz, T.; Schäffner, Zapf A.; Spannenberg, A.; Börner,
66. Cheng, Y.; Prusoff, W. H. Biochem. Pharmacol. 1973, 22, 3099.
A.; Beller, M. Chem. Eur. J. 2009, 15, 1329.