Functional reversal of (?)-Stepholidine analogues by replacement of benzazepine substructure using the ring-expansion strategy

Article abstract of DOI:10.1111/cbdd.12796

(?)-Stepholidine is an active ingredient of the Chinese herb Stephania and naturally occurring tetrahydroprotoberberine alkaloid with mixed dopamine receptor D₁ agonistic and dopamine receptor D₂ antagonistic activities. In this work, a series of novel hexahydrobenzo[4,5]azepino [2,1-a]isoquinolines were designed and synthesized as ring-expanded analogues of (?)-Stepholidine. Initial pharmacological assays demonstrated that a benzazepine replacement was associated with significant increase in selectivity and functional reversal at dopamine receptor D₁. Compound-(?)-15e (K_i?=?5.32?±?0.01?nm) is more potent than (?)-Stepholidine (K_i?=?13?nm) and was identified as a selective dopamine receptor D₁ antagonist (IC₅₀?=?0.14?μm). Moreover, molecular modeling suggested that (?)-15e might exert its dopamine receptor D₁ antagonistic activities through interacting with the transmembrane helix 7 of dopamine receptor D₁.

Full text of DOI:10.1111/cbdd.12796

Received: 24 February 2016
DOI: 10.1111/cbdd.12796
Revised: 10 May 2016
Accepted: 24 May 2016
|
|
R E S E A R C H A R T I C L E
Functional reversal of (−)-Stepholidine analogues by replacement
of benzazepine substructure using the ring-expansion strategy
Wei Li1,† | Li Zhang2,† | Lili Xu1,† | Congmin Yuan1 | Peng Du1 | Jiaojiao Chen2
Xuechu Zhen2 | Wei Fu1
|
1Department of Medicinal Chemistry & Key
Laboratory of Smart Drug Delivery, Ministry of
Education & PLA, School of Pharmacy, Fudan
University, Shanghai, China
2Jiangsu Key Laboratory of Translational Research
and Therapy for Neuropsychiatricdisorders
& Department of Pharmacology, College of
Pharmaceutical Sciences, Soochow University,
Suzhou, Jiangsu Province, China
(−)-Stepholidine is an active ingredient of the Chinese herb Stephania and natu-
rally occurring tetrahydroprotoberberine alkaloid with mixed dopamine receptor
D1 agonistic and dopamine receptor D2 antagonistic activities. In this work, a
series of novel hexahydrobenzo[4,5]azepino [2,1-a]isoquinolines were designed
and synthesized as ring-expanded analogues of (−)-Stepholidine. Initial pharma-
cological assays demonstrated that a benzazepine replacement was associated with
significant increase in selectivity and functional reversal at dopamine receptor
D1. Compound-(−)-15e (Ki = 5.32 0.01 nm) is more potent than (−)-Stepholidine
(Ki = 13 nm) and was identified as a selective dopamine receptor D1 antagonist
(IC50 = 0.14 μm). Moreover, molecular modeling suggested that (−)-15e might
exert its dopamine receptor D1 antagonistic activities through interacting with the
transmembrane helix 7 of dopamine receptor D1.
Correspondence
Wei Fu, wfu@fudan.edu.cn;
Xuechu Zhen, Zhenxuechu@suda.edu.cn
K E Y W O R D S
dopamine D1 receptor, Receptor and ligands (agonist/antagonist), selective D1R
antagonists
As an important monoamine neurotransmitter in the CNS,
dopamine plays an important role in reward-motivated
behavior,[1,2] mood regulatory functions,[3] learning mem-
ory, attention, and executive functions.[4] Dopamine system
dysfunction is associated with Parkinson’s disease,[5,6]
schizophrenia,[7–9] attention-deficit hyperactivity disor-
der,[10,11] and other CNS diseases. Among the five dopa-
mine receptor subtypes currently identified, the dopamine
receptor D1 (D1R) is most abundant. Although the clinical
utility of D1R antagonists in the treatment of schizophrenia
was limited,[12,13] these ligands were found to be effective
in the treatment of other rare CNS conditions, such as
pathological gambling[14] and Tourette’s syndrome[15] in
clinical trials. Therefore, discovery and development of
novel D1R antagonists might represent a field of research
interests for certain diseases of unmet medical need.
Unlike D1 agonists with enormous structural diversity,
most D1 antagonists are benzazepine derivatives.[16] SCH
23390[17] (Figure 1) was the first reported selective dopamine
D1 antagonist, with potent binding affinities at both D1 and
D5 dopamine receptors. This compound is now utilized as a
pharmacological tool to investigate the pathological roles of
dopamine D1 receptors in the development of CNS diseases
such as seizure, psychosis, and Parkinson’s disease.[18,19]
Another benzazepine-based D1 antagonist, ecopipam (SCH
39166)[20] (Figure 1), has ever been developed for the indi-
cation of treatment of psychotic disorders. Moreover, it
showed efficacy for obesity,[21] although its developabil-
ity as an antiobesity agent was limited due to the adverse
effects such as anxiety and depression.[22] Nevertheless, this
agent is now in clinical studies to support the indication of
Tourette’s syndrome in children (ClinicalTrials.gov Identi-
fier: NCT02102698).[15,23]
(−)-Stepholidine (l-SPD) (Figure 1) is a naturally occur-
ring tetrahydroprotoberberine (THPB) alkaloid isolated from
the Chinese herb Stephania intermedia.[24,25] It was identi-
fied as a unique dopaminergic ligand with dual D1R agonistic
†These authors contributed equally to this work.
|
Chem Biol Drug Des 2016; 1–9
wileyonlinelibrary.com/cbdd
© 2016 John Wiley & Sons A/S.
1

2
Li et aL.
|
FIGURE 1 Structure of benzazepines,
(−)-Stepholidine, and novel hexahydrobenzo[4,5]
azepino [2,1-a]isoquinolines as dopamine ligands
and dopamine receptor D2 (D2R) antagonistic activities. It
showed antipsychotic activities in animal models in vivo,[26]
as well as potential efficacy against opiate addiction[27] and
sleeping disorders.[28] Our prior efforts have been focused
on the molecular mechanism of l-SPD to confer multiple
pharmacological activities by in silico studies,[29,30] and more
l-SPD derivatives or analogues are expected to be available
not only to test the hypotheses we proposed in silico studies,
but also to provide drug candidates with potential clinical
utility.
implemented by DAICEL Daicel Chiral Technologies
(Shanghai, China) using preparative chiral HPLC (Column:
CHIRALCEL OJHs, Size: 0.46 cm I.D. × 15 cm L). The
optical rotation of chiral compounds was determined on
a Jasco (Easton, MD, USA) P-1020 Polarimeter.
1.2 | Synthesis procedures
The procedure for all intermediates and characterization was
described in the supporting information supplemented.
In this work, the six-membered piperidine substruc-
ture (Ring-C) of l-SPD (Figure 1) was expanded to seven-
membered benzazepine to yield hexahydrobenzo[4,5]azepino
[2,1-a]isoquinolines (Figure 1, 15–19), which hops over the
THPB scaffold of l-SPD. The effect of this scaffold hopping
on the pharmacological profiles of this series of derivatives
was examined at the molecular level. As a consequence,
unlike l-SPD that functions as a D1R agonist and D2R antag-
onist, some of hexahydrobenzo[4,5]azepino [2,1-a]isoquino-
lines were identified as potent and selective D1R antagonists,
suggesting the probable existence of functional activity cliff
between these structure categories.
1.3 | 2,3,11,12-tetramethoxy-5,6,8,9,14,14a-
hexahydrobenzo[4,5]azepino[2,1-a]isoquinoline
(15a)
This compound was prepared in the manner described
above for compound 15e using 14a to afford the title
1
compound in 70.1% yield as an off-white solid: H NMR
(400 MHz, CDCl3) δ 6.75 (s, 1H), 6.71 (s, 1H), 6.66 (s,
1H), 6.58 (s, 1H), 3.91–3.86 (m, 13H), 3.54 (dd, J = 15.7,
9.8 Hz, 1H), 3.31–3.59 (m, 3H), 3.01–2.89 (m, 2H),
2.81–2.74 (m, 3H), 2.58 (dd, J = 15.3, 5.9 Hz, 1H).
ESI-MS m/z 370.3 [M + H]+.
1
|
MATERIALS AND METHODS
1.4 | 3,11,12-trimethoxy-5,6,8,9,14,14a-
hexahydrobenzo[4,5]azepino[2,1-a]isoquinolin-
2-ol (15b)
1.1 | Chemistry
Commercially available chemicals were used directly without
further purification. Moisture-sensitive reactions were imple-
mented under an atmosphere of dry argon, and relevant
solvents were dried according to standard procedures.
Reactions were monitored using thin-layer chromatography
on silica-coated plastic sheets (silica gel 60F254; Rushan
Taiyang Desiccant Co., Ltd., Rushan, China) with the indi-
cated eluant. Nuclear magnetic resonance spectra were
recorded on a Brucker-DPX 400 MHz spectrometer (Bruker,
Scientific Technology Co., Ltd., Beijing, China), and mass
spectral data were collected on a HP5973N analytical mass
spectrometer. Separation works of chiral compounds were
This compound was prepared in the method described
above for compound 15e using 14a to afford the title
compound in 52.6% yield as a light yellow solid: 1H
NMR (400 MHz, CDCl3) δ 6.84 (s, 1H), 6.77 (s, 1H),
6.64 (s, 1H), 6.55 (s, 1H), 3.91 (s, 3H), 3.87 (s, 3H),3.86
(s, 3H), 3.78 (d, J = 9.3 Hz, 1H), 3.47 (dd, J = 15.3,
9.5 Hz, 1H), 3.33–3.19 (m, 3H), 3.00–2.70 (m, 5H),
2.59 (dd, J = 14.5, 5.4 Hz, 1H).13C NMR (100 MHz,
CDCl3) δ 147.0, 147.0, 145.1, 143.8, 134.1, 133.3, 131.9,
126.3, 113.4, 113.1, 113.0, 110.5, 63.3, 57.5, 56.1, 56.1,
55.9, 49.2, 42.9, 34.5, 29.3. ESI-MS m/z 356.2
[M + H]+.

Li et aL.
3
|
(100 MHz, CDCl3) δ 145.3, 145.1, 143.9, 142.7, 127.1,
125.6, 120.0, 113.0, 110.4, 107.9, 63.0, 56.1, 55.9, 53.4,
49.2, 31.9, 29.7, 29.4, 24.2, 22.7, 14.1. ESI-MS m/z
342.1 [M + H]+.
1.5 | 2,3,10,11-tetramethoxy-5,6,8,9,14,14a-
hexahydrobenzo[4,5]azepino[2,1-a]isoquinoline
(15c)
This compound was prepared in the manner described
above for compound 15e using 14c to afford the title
1
compound in 81.3% yield as an off-white solid: H NMR
1.8 | 3,11-dimethoxy-5,6,8,9,14,14a-
hexahydrobenzo[4,5]azepino[2,1-a]
isoquinoline-2,10-diol (−)-15e
(400 MHz, CDCl3) δ 6.95 (d, J = 8.2 Hz, 1H), 6.70
(d, J = 8.0 Hz, 2H), 6.57 (s, 1H), 3.89 (s, 3H), 3.86
(d, J = 2.5 Hz, 7H), 3.81 (d, J = 9.5 Hz, 1H), 3.76
(s, 3H), 3.51 (dd, J = 15.1, 9.5 Hz, 1H), 3.32–3.18
(m, 3H), 3.02–2.72 (m, 6H).13C NMR (100 MHz, CDCl3)
δ 25.0, 29.1, 42.6, 48.2, 55.7, 55.8, 56.0, 56.8, 61.0,
63.0, 109.1, 110.2, 111.0, 124.3, 126.9, 131.1, 134.7,
136.1, 146.3, 147.1, 147.4, 151.2. ESI-MS m/z 370.1
[M + H]+.
This compound was separated by preparative chiral HPLC
(Column: CHIRALCEL OJHs, Size: 0.46 cm
I.D. × 15 cm L MeOH/CH3CN/DEA = 90/10/0.1 (v/v/v),
Flow rate:1.0 mL/min, Wave length: UV 254 nm) from
15e as an white solid: RT = 5.057 min, e.e. > 99%,
1
[ꢀ]24 = −3.377° (c = 1.14, DMSO); H NMR (400 MHz,
D
DMSO) δ 8.69 (s, 1H), 8.35 (s, 1H), 6.67–6.62 (m,
3H), 6.55 (s, 1H), 3.74 (s, 3H), 3.69 (s, 3H), 3.43–3.36
(m, 2H), 3.27–3.21 (m, 1H), 3.14–3.07 (m, 3H), 2.76–2.67
(m, 3H), 2.57–2.46 (m, 2H). ESI-MS m/z 342.0
[M + H]+.
1.6 | 3,10,11-trimethoxy-5,6,8,9,14,14a-
hexahydrobenzo[4,5]azepino[2,1-a]isoquinolin-
2-ol (15d)
This compound was prepared in the manner described
above for compound 15e using 14d to afford the title
1.9 | 3,11-dimethoxy-5,6,8,9,14,14a-
hexahydrobenzo[4,5]azepino[2,1-a]
isoquinoline-2,10-diol (+)-15e
1
compound in 85.3% yield as an off-white solid: H NMR
(400 MHz, CDCl3) δ 6.95 (d, J = 8.2 Hz, 1H), 6.81 (s,
1H), 6.71 (d, J = 8.0 Hz,1H), 6.55 (s, 1H), 3.86 (s, 3H),
3.85 (s, 3H),3.76 (s, 3H), 3.72 (t, J = 5.1 Hz, 1H), 3.45
(dd, J = 15.3, 9.4 Hz, 1H), 3.29–3.21 (m, 3H), 3.00–2.68
(m, 6H). 13C NMR (100 MHz, CDCl3) δ 151.3, 146.4,
145.1, 143.8, 136.0, 134.9, 132.0, 126.3, 124.5, 113.0,
110.5, 110.0, 63.4, 61.0, 57.2, 56.0, 57.8, 49.2, 42.9, 29.3,
25.4. ESI-MS m/z 356.2 [M + H]+.
This compound was separated by preparative chiral HPLC
(Column: CHIRALCEL OJHs, Size: 0.46 cm I.D. ×
15 cm L MeOH/CH3CN/DEA = 90/10/0.1 (v/v/v), Flow
rate: 1.0 mL/min, Wave length: UV 254 nm) from 15e
as an white solid: RT = 3.452 min, e.e. > 99%, 1H
NMR (400 MHz, DMSO) δ 8.69 (s, 1H), 8.35 (s, 1H),
6.67–6.62 (m, 3H), 6.55 (s, 1H), 3.74 (s, 3H), 3.69 (s,
3H), 3.43–3.36 (m, 2H), 3.27–3.21 (m, 1H), 3.14–3.07
(m, 3H), 2.76–2.67 (m, 3H), 2.57–2.46 (m, 2H). ESI-MS
m/z 342.0 [M + H]+.
1.7 | 3,11-dimethoxy-5,6,8,9,14,14a-
hexahydrobenzo[4,5]azepino[2,1-a]
isoquinoline-2,10-diol (15e)
To a solution of LiAlH4 (0.2 g, 5.27 mmol) suspended
in dry THF (30 mL) was added 14e (0.4 g, 1.13 mmol)
in THF (10 mL) at 80 °C. The mixture was stirred
for 3 h, and then, H2O (1.9 mL) and 15% aqueous
HCl (1.9 mL) were added successively. The reaction
mixture was concentrated, and water (50 mL) was added
and extracted with ethyl acetate (3 × 50 mL).The organic
extract was washed three times with brine, dried over
Na2SO4, filtered, and the solvent was removed. The
resulting residue was purified by column chromatography,
eluting with CH2Cl2: MeOH=20:1 to yield 0.11 g (28.6%)
1.10 | 5,6,8,9,14,14a-hexahydrobenzo[4,5]
azepino[2,1-a]isoquinoline-2,3,10,11-tetraol (18)
A solution of 15a (50 mg, 0.14 mmol) in 48%HBr was
stirred at 90 °C for 7 h. The mixture was then quenched
with saturated NaHCO3 and extracted with EtOAc
(3 × 20 mL). The organic layer was dried over anhy-
drous Na2SO4 filtered and purified by chromatography
(DCM:MeOH = 20:1) to get the desired product as a
light yellow solid (23 mg, 41.8%):1H NMR (400 MHz,
DMSO) δ 9.05 (s, 1H), 8.67 (s, 1H), 8.59 (s, 1H),
8.03 (s, 1H), 6.57 (s, 1H), 6.48 (d, J = 3.4 Hz, 2H),
6.37 (s, 1H), 4.01 (d, J = 7.2 Hz, 1H), 3.72 (d,
J = 17.8 Hz, 1H), 3.19 (dd, J = 14.3, 9.1 Hz, 2H),
3.22–3.05 (m, 3H), 2.71–2.62 (m, 4H).ESI-MS m/z 314.2
[M + H]+.
1
of 15e as an off-white solid: H NMR (400 MHz, CDCl3)
δ 6.81 (s, 1H), 6.74 (d, J = 8.1 Hz, 1H), 6.66 (d,
J = 8.1 Hz, 1H), 6.54 (s, 1H), 3.87–3.85 (m, 7H),
3.48 (dd, J = 15.3, 9.2 Hz, 1H), 3.36–3.19 (m, 3H),
3.05–2.89 (m, 4H), 2.85–2.72 (m, 2H). 13C NMR

4
Li et aL.
|
tetraacetic acid (EDTA), 100 mm NaCl, 1 mm DL-
dithiothreitol, and 40 μm guanosine triphosphate. The reac-
tion was initiated by the addition of [35S]GTPγS (final
concentration of 0.1 nm). Non-specific binding was measured
in the presence of 100 μm 50-guanylimidodiphosphate
(Gpp(NH)p). The reaction was terminated by the addition
of 1 mL of ice-cold washing buffer (50 mm Tris, pH 7.5,
5 mm MgCl2, 1 mm EDTA, and 100 mm NaCl) and was
rapidly filtered with GF/C glass fiber filters (Whatman)
and washed three times. Radioactivity was determined by
liquid scintillation counting. Again, l-SPD was used as the
control of D1R agonist.[31,32]
1.11 | 3,10,11-trimethoxy-5,6,8,9,14,14a-
hexahydrobenzo[4,5]azepino[2,1-a]isoquinolin-
2-yl acetate (19)
To a solution of 15d (40 mg, 0.11 mmol), TEA (24 μL,
0.17 mmol) in DCM (10 mL) was added acetylchloride
(12 μL, 0.11 mmol) and DMAP (13 mg, 0.11 mmol) at
0 °C, then stirred at room temperature for 4 h. The reac-
tion mixture was poured into water (20 mL) and extracted
with DCM (3 × 15 mL). The organic layer was washed
with water (3 × 15 mL), dried over anhydrous Na2SO4,
and filtered, and the solvent was removed under vacuum.
Chromatography (DCM:MeOH = 50:1) afforded 35 mg
1
(80.1%) of the desired product as a light yellow solid: H
1.12.3 | Molecular modeling
NMR (400 MHz, CDCl3) δ 6.94 (d, J = 8.2 Hz, 1H), 6.92
(s, 1H), 6.70 (d, J = 8.2 Hz, 1H), 6.66 (s, 1H), 3.86 (s,
3H), 3.81 (s, 3H), 3.76 (s, 3H), 3.70 (d, J = 9.3 Hz, 1H),
3.43 (dd, J = 15.1, 9.5 Hz, 1H), 3.28–3.20 (m, 3H), 3.03–2.69
(m, 6H), 2.33 (s, 3H).13C NMR (100 MHz, CDCl3) δ 169.2,
151.3, 149.1, 146.4, 138.0, 136.0, 134.7, 133.6, 131.8, 124.5,
121.2, 112.2, 109.5, 63.3, 61.0, 57.4, 55.9, 55.8, 49.2, 43.2,
29.8, 25.6, 20.7. ESI-MS m/z 398.2 [M + H]+.
The active and inactive structures of D1R were built by
homology modeling module encoded in Discovery Studio
3.5[33] by taking the crystal structures of active-state and
inactive-state β2AR as templates (PDB codes: 3SN6 and
2RIH). The sequence alignment of D1R and β2AR is
shown in Figure S1 (Supporting information), and the
sequence identity and similarity is 38.5% and 63.8% between
D1R and β2AR, respectively. Fifty models for both active
and inactive structures of D1R were generated after loop
refinement and the one with the lowest discrete optimized
protein energy score was submitted to the stepwise energy
minimization to get the structure of D1R.
1.12 | Biological evaluations
1.12.1 | Binding assays
Molecular docking of l-SLR and (−)-15e (constructed
and energy minimized in vacuum using SYBYL 6.9 with
the MMFF94s force field) in the binding site of D1R was
performed using goLdsuite 5.0 (The Cambridge Crystallo-
graphic Data Centre (CCDC), Cambridge, UK). The bind-
ing pocket was defined as amino acids within 15 Å of the
Cγ carbon atom of D3.32. Ten conformations were produced
for each ligand. Other parameters were set as default. Thirty
docking orientations were obtained. The final docking pose
was chosen based on the GOLD score and the predicted
binding energy; in the meantime, the reported binding mode
between l-SPD and D1 receptor and the results of mutation
experiments on these receptors were also considered.
All target molecules were profiled by the competitive
binding assays for D1R, D2R, and D3R, respectively, using
a membrane preparation harvested from stably transfected
HEK293 cells. The transfection of plasmid and membrane
preparations was conducted as described in our previous
report, with 0.7 nm [3H] SCH23390 (D1R) or [3H]Spiperone
(D2R and D3R) as the standard radioligands. 1 nm–10 μm
l-SPD and l-SLR were used as the control for the assay.
For the preliminary screening, the potential of each com-
pound at a concentration of 10 μm to inhibit the binding
of a tritiated-labeled ligand to its receptor was examined.
For compounds with the percent inhibition greater than
90%, Ki/IC50 was determined. Ki was calculated based on
the following formula: Ki = IC50/(1 + C/Kd).
2
|
RESULTS
1.12.2 | Functional assays
The membrane protein of HEK cells stably transfected
with D1 receptor was used.[31] The [35S]GTPγS binding
assay was performed at 30 °C for 30 min containing 10 μg
of membrane protein in a final volume of 100 μL with
various concentration of the compound. The antagonism
effects of the compounds were tested in the existence of
10 μm haloperidol for the D1R. The binding buffer contains
50 mm Tris, pH 7.5, 5 mm MgCl2, 1 mm ethylenediamine
2.1 | Chemistry
As depicted in retro-synthetic analysis (Scheme 1), the
target molecule 15a–e could be disconnected into the key
intermediate 13a–e, which could be further decomposed
into two critical fragment 6a–e and 12a–e. With com-
mercially available vanillin and nitromethane as the starting
material, the critical fragment 6a–b was prepared

Li et aL.
5
|
SCHEME 1 Retrosynthetic analysis of compounds
15–19
SCHEME 2 Synthesis of the key fragment 6a–b.
Reagents and conditions: (a) BnBr, K2CO3, MeOH;
(b) CH3NO2, NH4OAc, AcOH; (c) KBH4, THF;
(d) NaNO2, AcOH, DMSO; (e) LiAlH4, THF; (f) TsCl,
TEA, DCM; (g) NaI, acetone
SCHEME 3 Synthesis of the key fragment 12a.
Reagents and conditions: (a) BnBr,K2CO3; (b) CH3NO2,
NH4OAc, AcOH; (c) KBH4; (d) NaNO2, AcOH, DMSO;
(e) aminoacetaldehyde dimethyl acetal, EDC, HOBt,
TEA; (f) HCl, HOAc; (g) BnBr, K2CO3
SCHEME 4 Synthesis of 15a–15e. Reagents
and conditions: (a) tBuOK, DMF; (b) HCl/EtOH;
(c) LiAlH4, THF
subsequently via Henry condensation, selective olefinic
reduction, nitro group oxidation into carboxylic acids,
reduction of carboxylate, and iodide substitution, as depicted
in Scheme 2. Meanwhile, with o-vanillin as the starting
material, the key intermediate 12a was synthesized via
alkylation, aldol-like condensation with nitromethane, selec-
tive olefinic reduction, nitro group oxidation into carboxylic
acids, amidation, cyclization, and alkylation again[34]
(Scheme 3). Furthermore, the other key fragments 12b
and 12c (Scheme 4) were prepared via similar conditions,
condensed to afford the key intermediate 13a–e, which
was subsequently cyclized under acidic conditions to pro-
duce the expected molecule 15a–e (Scheme 4).
In addition, compound 16, a hexahydrobenzo[4,5]aze-
pino [2,1-a]isoquinoline, as well as its demethylated deriv-
ative compound 17 was prepared according to the procedure
described in the literature35 (Scheme 5). Compound 15c
was demethylated to produce compound 18 (Scheme 6).
Compound 15d was treated with acetyl chloride to afford
the acylate compound 19 (Scheme 7). As to compound 15e,
the racemic mixture was resolved using preparative chiral
HPLC to afford its optical isomers, for example, (+)-15e and
(−)-15e.
with
2-(3,4-dimethoxyphenyl)
acetic
acid
and
2-(2,3-dimethoxyphenyl) acetic acid as the corresponding
starting material. Intermediate 6a–b and 12a–c were

6
Li et aL.
|
SCHEME 5 Synthesis of 16 and 17. Reagents and
conditions: (a) NaBH4, MeOH; (b) BrCH2CN; (c) NaH,
THF; (d) NaBH4, EtOH; (e) BCl3, DCM
TABLE 1 Binding affinities of novel hexahydrobenzo[4,5]azepino
[2,1-a]isoquinolines for dopamine receptors D1R, D2R, and D3R
Inhibition % or Ki ( SEM, nm)
Compound
15a
D1R
D2R
D3R
13.32
33.94
26.53
15b
69.09 8.04
57.27
6.96
29.82
83.95
36.59
4.26
39.22
21.97
SCHEME 6 Synthesis of compound 18. Reagents and conditions: (a)
48% HBr, reflux
15c
15d
30.28 2.13
8.80 0.44
60.32
284.34 7.45
88.63
15e
16
27.28
17
220.27 23.71
33.74 2.36
474.10 19.30
72.76
39.24
72.15
51.97
34.40
58.26
85
61.20
18
470.86 15.58
73.67
19
(+)-15e
(−)-15e
l-SPDa
l-SLRa
SCH23390
Spiperone
19.85
5.32 0.01
13
1185.20 43.31
–
SCHEME 7 Synthesis of compound 19. Reagents and conditions:
(a) CH3COCl, TEA, DCM
42
180
–
1.69 0.12
ND
ND
ND
2.2 | Biological evaluation
1.08 0.09
0.48 0.07
Results of percent inhibition or Ki values and functional
ND, not determined.
assays are summarized in Table
respectively.
1 and Table 2,
aThe binding affinities of l-SPD and l-SLR were cited from Xu’s paper.[37]
TABLE 2 The [35S]GTPγS binding assays of compound-15d, (−)-15e,
and 18 for D1R
3
|
DISCUSSION
D1R
Compound
15d
EC50 (μm)
IC50 (μm)
3.1 | Binding assays
a
–
33.45
0.14
114.50
–
Compared with the prototype compound l-SPD with THPBs
scaffold, the ring-expanded analogues generally demonstrated
loss in D2R binding affinities. However, the D1R binding
affinities were maintained or minimally decreased for analogues
with 2-hydroxy substitution, resulting in significant increase
in D1R selectivity. In fact, the structure motif of m-hydroxy
phenethylamine motif (as for l-SPD and its analogues,
2-hydroxy substitution) was considered to be the conserved
substructure for ligands with dopamine agonistic activities,
which was also supported by the significant loss of binding
affinities in D1R for the analogues with 2-methoxy (compound-
15a, compound-15c) or 2,3-dioxole (compound-16) without
such substitution. In contrast, 3-hydroxy substitution also
contributed to D1R binding affinities, although there were
no substantial differences in D1R binding affinities between
(−)15e
18
–
–
l-SPD
41.10b
SCH23390
SKF-38393
–
0.52
0.14
[
a 35S] GTPγS binding activity could not be detected.
bEC50 value of l-SPD was cited from Dong et al.’s paper.[32]
compound-17 (220.27
23.71 nm) with 3-hydroxy substitu-
tion versus compound-15d (30.28 2.13 nm) with 3-methoxy
substitution. In addition, compound-15e (8.80 0.44 nm) with
10-hydroxy substitution demonstrated relatively higher D1R
affinities than that of compound-15d (30.28 2.13 nm) with
10-methoxy substitution, suggesting 10-hydroxy substitution

Li et aL.
7
|
FIGURE 2 The chemical structures of related l-SPD
analogues
might also contribute to D1R affinities. As to the optimal
at the D1 receptor (Figure 2), resulting in the antagonistic
activities against D1R. As the benzazepine replacement
changed, the spatial orientation of the D-ring and all com-
pounds with the benzazepine replacement assayed showed
antagonistic activities against D1R, suggesting the disparity
in the mechanism of functional reversal at D1R for the
ring-expanded analogues of l-SPD might be different.
Moreover, the substituents similar to those described by
Qian et al.[36] could only regulate the extent of antagonistic
potency other than the nature of biological function, indi-
cating the correlation between the benzazepine replacement
and the D1R antagonistic nature for l-SPD analogues.
compound-15e, its (−)-isomer [(−)-15e] (5.32
0.01 nm)
showed much more potent binding affinities at D1R compared
with its (+)-isomer [(+)-15e] (72.76% @ 10 μm), suggesting
certain stereochemical requirement for these analogues to
maintain favorable D1R binding affinities.
As for D3R binding affinities, only compound-15, com-
pound-18, and compound-15e with 2-hydroxy substitution
demonstrated mild-to-moderate D3R binding affinities, suggest-
ing a role of this substituent conferring D3R binding activities.
3.2 | Functional assays
The optimal compounds identified in binding assays were
further examined by functional assays to evaluate their
biological functions (Table 2). In contrast to l-SPD with
no intrinsic antagonistic activities on D1R, 15d, (−)-15e,
and 18 with the benzazepine replacement generally dem-
onstrated pure D1R antagonistic activities. Moreover, in
comparison with SCH23390 (IC50 = 0.52 μm), (−)-15e
(IC50 = 0.14 μm) showed comparable or even higher
antagonistic activities for D1R. According to the work
depicted by Qian et al.[36] removal of C10 hydroxy group
and introduction of a methoxy group at C11 of the phar-
macophore of l-SPD can reverse the function of THPB s
3.3 | Molecular modeling
The effect of benzazepine replacement on the binding
and function at D1R was further investigated for com-
pound (−)-15e. Due to the similarities in the spatial ori-
entation of ring-C and relevant substituents, l-Scoulerine
(l-SLR, Figure 2), another naturally occurring product with
similar binding profiles at dopamine receptors[37] was
chosen as the reference compound for molecular modeling.
l-SLR and (−)-15e were docked to the D1R by using
goLdsuite 5.0[38] (Figure 3). The active and inactive struc-
tures of D1R were built by homology modeling, taking
FIGURE 3 Predicted Binding mode of l-SLR (a)
and (−)-15e (b) with D1R

8
Li et aL.
|
the crystal structures of active-state and inactive-state β2AR
as templates (PDB ID: 3SN6[39] and 2RH1[40]). Based
on the results of molecular docking, (−)-15e took a reverse
docking pose in the active site of (−)-15e-D1R complex
compared with l-SLR. In both models, the salt bridge
between the protonated nitrogen atom of ligand and D3.32
was maintained, consistent with the fact that mutation of
D3.32 is associated with the binding affinities of both
antagonist and agonist among catecholamine receptors.[41,42]
In the l-SLR-D1R complex, the 2-hydroxy group of l-
SLR formed a hydrogen bond with S5.42. But in
(−)-15e-D1R complex, the residue S5.42 interacted with
the 10-hydroxy group of ligand via a hydrogen bonding
interaction. Furthermore, in comparison with the rigid and
flat THPB substructure of l-SLR, compound (−)-15e could
take more flexible and twisted conformations as the con-
sequence of benzazepine replacement. In addition, according
to the docked model of (−)-15e-D1R complex, ring-A of
ligand was extended to the transmembrane VII and the
10-hydroxy group formed another hydrogen bond with
D7.36, indicating the possible role of TM7 in the antago-
nistic binding of compound-(−)-15e.
REFERENCES
[1] S. Karunakaran, A. Chowdhury, F. Donato, C. Quairiaux, C. M. Michel,
P. Caroni. Nat. Neurosci. 2016, 19, 454.
[2] E. A. Ferenczi, K. A. Zalocusky, C. Liston, L. Grosenick, M. R. Warden, D.
Amatya, K. Katovich, H. Mehta, B. Patenaude, C. Ramakrishnan, P. Kala-
nithi, A. Etkin, B. Knutson, G. H. Glover, K. Deisseroth. Science 2016, 351,
aac9698.
[3] H. Q. Yan, N. D. Osier, J. Korpon, J. W. Bales, A. E. Kline, A. K. Wagner,
C. E. Dixon. In Persistent Cognitive Deficits: Implications of Altered Dopa-
mine in Traumatic Brain Injury (Ed. F. H. Kobeissy). CRC Press/Taylor &
Francis, Boca Raton, FL 2015.
[4] C. L. German, M. G. Baladi, L. M. McFadden, G. R. Hanson, A. E. Flecken-
stein. Pharmacol. Rev. 2015, 67, 1005.
[5] R. M. Zweig, E. A. Disbrow, V. Javalkar. Neurol. Clin. 2016, 34, 235.
[6] P. T. Bell, M. Gilat, C. O’Callaghan, D. A. Copland, M. J. Frank, S. J.
Lewis, J. M. Shine. Hum. Brain Mapp. 2015, 36, 1278.
[7] Y. Yamaguchi, Y. A. Lee, Y. Goto. Front Neurosci 2015, 9, 219.
[8] J. Lally, J. H. MacCabe. Br. Med. Bull. 2015, 114, 169.
[9] T. Krzymowski, S. Stefanczyk-Krzymowska. J. Physiol. Pharmacol. 2015,
66, 331.
[10] J. Wu, H. Xiao, H. Sun, L. Zou, L. Q. Zhu. Mol. Neurobiol. 2012, 45, 605.
[11] N. D. Volkow, G. J. Wang, J. H. Newcorn, S. H. Kollins, T. L. Wigal, F.
Telang, J. S. Fowler, R. Z. Goldstein, N. Klein, J. Logan, C. Wong, J. M.
Swanson. Mol. Psychiatry 2011, 16, 1147.
[12] J. A. Den Boer, H. J. van Megen, W. W. Fleischhacker, J. W. Louwerens, B.
R. Slaap, H. G. Westenberg, G. D. Burrows, O. N. Srivastava. Psychophar-
macology 1995, 121, 317.
[13] P. Karlsson, L. Smith, L. Farde, C. Harnryd, G. Sedvall, F. A. Wiesel. Psy-
chopharmacology 1995, 121, 309.
[14] J. E. Grant, B. L. Odlaug, D. W. Black, T. Fong, M. Davtian, R. Chipkin, S.
W. Kim. Ann. Clin. Psychiatry 2014, 26, 179.
4
|
CONCLUSIONS
[15] D. L. Gilbert, C. L. Budman, H. S. Singer, R. Kurlan, R. E. Chipkin. Clin.
Neuropharmacol. 2014, 37, 26.
In conclusion, as the ring-expanded analogues of l-SPD,
a series of novel hexahydrobenzo[4,5]azepino [2,1-a]iso-
quinolines were designed, synthesized, and assayed. As
demonstrated by preliminary SAR investigations, the ben-
zazepine replacement was associated with significant
increase in selectivity of D1R and its functional reversal.
Compound (−)-15e was identified as a potent and selec-
tive D1R antagonist. This was in comparison with l-SPD
with dual D1R agonistic and D2R antagonistic activities.
Moreover, molecular modeling suggested (−)-15e might
exert its D1R antagonistic activities through interacting
with the TM7 of D1R.
[16] J. Zhang, B. Xiong, X. Zhen, A. Zhang. Med. Res. Rev. 2009, 29, 272.
[17] D. W. Schulz, E. J. Stanford, S. W. Wyrick, R. B. Mailman. J. Neurochem.
1985, 45, 1601.
[18] L. C. Iorio, A. Barnett, F. H. Leitz, V. P. Houser, C. A. Korduba. J. Pharma-
col. Exp. Ther. 1983, 226, 462.
[19] Y. Itoh, M. Beaulieu, J. W. Kebabian. Eur. J. Pharmacol. 1984, 100, 119.
[20] R. E. Chipkin, L. C. Iorio, V. L. Coffin, R. D. McQuade, J. G. Berger, A.
Barnett. J. Pharmacol. Exp. Ther. 1988, 247, 1093.
[21] A. Astrup, F. L. Greenway, W. Ling, L. Pedicone, J. Lachowicz, C. D.
Strader, R. Kwan, Ecopipam Obesity Study, Group. Obesity 2007, 15, 1717.
[22] P. J. Nathan, B. V. O’Neill, A. Napolitano, E. T. Bullmore. CNS Neurosci.
Ther. 2011, 17, 490.
[23] D. R. Shprecher, L. Schrock, M. Himle. Curr. Opin. Neurol. 2014, 27, 484.
[24] G. Z. Jin, K. X. Huang, B. C. Sun. Neurochem. Int. 1992, 20(Suppl), 175S.
[25] G. Z. Jin, Z. T. Zhu, Y. Fu. Trends Pharmacol. Sci. 2002, 23, 4.
[26] J. Mo, Y. Guo, Y. S. Yang, J. S. Shen, G. Z. Jin, X. Zhen. Curr. Med. Chem.
2007, 14, 2996.
[27] K. Yue, B. Ma, L. Chen, X. Tian, Q. Ru, Y. Gan, D. Wang, G. Jin, C. Li.
NeuroReport 2014, 25, 7.
[28] M. H. Qiu, W. M. Qu, X. H. Xu, M. M. Yan, Y. Urade, Z. L. Huang. Phar-
macol. Biochem. Behav. 2009, 94, 16.
ACKNOWLEDGMENTS
This work was supported by the National Natural Science
Foundation of China (No. 81473136, 31400624), State
Key
Laboratory
of
Toxicology
and
Medical
[29] B. Li, W. Li, P. Du, K. Q. Yu, W. Fu. J Phys Chem B 2012, 116, 8121.
[30] W. Fu, J. Shen, X. Luo, W. Zhu, J. Cheng, K. Yu, J. M. Briggs, G. Jin, K.
Chen, H. Jiang. Biophys. J . 2007, 93, 1431.
[31] Z. Li, J. Huang, H. Sun, S. Zhou, L. Guo, Y. Zhou, X. Zhen, H. Liu. Bioorg.
Med. Chem. 2014, 22, 5838.
Countermeasures (Academy of Military Medical Science),
No.TMC201501 and Shanghai Science and Technology
Development Funds (14431900500).
[32] Z. J. Dong, X. Guo, L. J. Chen, Y. F. Han, G. Z. Jin. Life Sci. 1997, 61, 465.
[33] A. S. Inc, Discovery Studio Modeling Environment, Accelrys Software Inc,
San Diego 2012.
CONFLICT OF INTEREST
[34] M. Reiffen, W. Eberlein, P. Muller, M. Psiorz, K. Noll, J. Heider, C. Lillie,
W. Kobinger, P. Luger. J. Med. Chem. 1990, 33, 1496.
All authors declared no conflicts of interest.

Li et aL.
9
|
[35] Y. X. Liu, S. F. Chen, X. T. Liang. Chin. Chem. Lett. 1999, 10, 637.
[36] W. Qian, W. Lu, H. Sun, Z. Li, L. Zhu, R. Zhao, L. Zhang, S. Zhou, Y. Zhou,
H. Jiang, X. Zhen, H. Liu. Bioorg. Med. Chem. 2012, 20, 4862.
[37] S. X. Xu, L. P. Yu, Y. R. Han, Y. Chen, G. Z. Jin. Acta Pharmacol. Sin.
1989, 10, 104.
[38] G. Jones, P. Willett, R. C. Glen, A. R. Leach, R. Taylor. J. Mol. Biol. 1997,
267, 727.
[41] M. F. Hibert, S. Trumpp-Kallmeyer, A. Bruinvels, J. Hoflack. Mol. Pharma-
col. 1991, 40, 8.
[42] S. Kortagere, S. Y. Cheng, T. Antonio, J. Zhen, M. E. Reith, A. K. Dutta.
Biochem. Pharmacol. 2011, 81, 157.
[39] S. G. Rasmussen, B. T. DeVree, Y. Zou, A. C. Kruse, K. Y. Chung, T. S.
Kobilka, F. S. Thian, P. S. Chae, E. Pardon, D. Calinski, J. M. Mathiesen, S.
T. A. Shah, J. A. Lyons, M. Caffrey, S. H. Gellman, J. Steyaert, G. Skiniotis,
W. I. Weis, R. K. Sunahara, B. K. Kobilka. Nature 2011, 477, 549.
[40] V. Cherezov, D. M. Rosenbaum, M. A. Hanson, S. G. Rasmussen, F. S.
Thian, T. S. Kobilka, H. J. Choi, P. Kuhn, W. I. Weis, B. K. Kobilka, R. C.
Stevens. Science 2007, 318, 1258.
SUPPORTING INFORMATION
Additional Supporting Information may be found online in
the supporting information tab for this article.