Design and synthesis of furyl/thineyl pyrroloquinolones based on natural alkaloid perlolyrine, lead to the discovery of potent and selective PDE5 inhibitors

Article abstract of DOI:10.1016/j.ejmech.2018.02.039

Based on perlolyrine (1), a natural alkaloid with weak PDE5 potency from the traditional Chinese aphrodisiac plant Tribulus terrestris L., a series α-substituted tetrahydro-β-carboline (THβC) derivatives were synthesized via T⁺BF4^--mediated oxidative C–H functionalization of N-aryl THβCs with diverse potassium trifluoroborates. Following Winterfeldt oxidation afforded the corresponding furyl/thienyl pyrroloquinolones, of which 5-ethylthiophene/ethylfuran derivatives 20a–b were identified as the most potent and selective PDE5 inhibitors. Among the enantiomers, (S)-20a and (S)-20b (IC₅₀ = 0.52 and 0.39 nM) were found to be more effective than their (R)-antipode, display favorable pharmacokinetic profiles, exert in vitro vasorelaxant effects on the isolated thoracic aorta, and exhibit in vivo efficacy in the anesthetized rabbit erectile model.

Full text of DOI:10.1016/j.ejmech.2018.02.039

Accepted Manuscript
Design and synthesis of furyl/thineyl pyrroloquinolones based on natural alkaloid
perlolyrine, lead to the discovery of potent and selective PDE5 inhibitors
Hongbo Zheng, Lin Li, Bin Sun, Yun Gao, Wei Song, Xiaoyu Zhao, Yanhui Gao, Zhiyu
Xie, Nianzhao Zhang, Jianbo Ji, Huiqing Yuan, Hongxiang Lou
PII:
S0223-5234(18)30169-7
10.1016/j.ejmech.2018.02.039
EJMECH 10216
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 4 September 2017
Revised Date: 12 February 2018
Accepted Date: 12 February 2018
Please cite this article as: H. Zheng, L. Li, B. Sun, Y. Gao, W. Song, X. Zhao, Y. Gao, Z. Xie, N. Zhang,
J. Ji, H. Yuan, H. Lou, Design and synthesis of furyl/thineyl pyrroloquinolones based on natural alkaloid
perlolyrine, lead to the discovery of potent and selective PDE5 inhibitors, European Journal of Medicinal
Chemistry (2018), doi: 10.1016/j.ejmech.2018.02.039.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

Graphic Abstract
ACCEPTED MANUSCRIPT

Eur. J. Med. Chem.
ACCEPTED MANUSCRIPT
Design and Synthesis of Furyl/Thineyl Pyrroloquinolones Based on Natural
Alkaloid Perlolyrine, Lead to the Discovery of Potent and Selective PDE5
Inhibitors
Hongbo Zheng,a Lin Li,a Bin Sun,a,b Yun Gao,a Wei Song,a Xiaoyu Zhao,a Yanhui Gao,a Zhiyu Xie,a Nianzhao
Zhang,a Jianbo Ji,a Huiqing Yuan,c and Hongxiang Loua *
aDepartment of Natural Products Chemistry, Key Laboratory of Natural Products & Chemical Biology,
Ministry of Education, School of Pharmaceutical Sciences, bNational Glycoengineering Research Center and
cDepartment of Biochemistry and Molecular Biology, School of Medicine, Shandong University, Jinan,
250012, China.
Abstract
Based on perlolyrine (1), a natural alkaloid with weak PDE5 potency from the traditional Chinese
aphrodisiac plant Tribulus terrestris L., a series α-substituted tetrahydro-β-carboline (THβC)
derivatives were synthesized via T+BF4--mediated oxidative C–H functionalization of N-aryl THβCs
with diverse potassium trifluoroborates. Following Winterfeldt oxidation afforded the corresponding
furyl/thienyl pyrroloquinolones, of which 5-ethylthiophene/ethylfuran derivatives 20a–b were
identified as the most potent and selective PDE5 inhibitors. Among the enantiomers, (S)-20a and
(S)-20b (IC50 = 0.52 and 0.39 nM) were found to be more effective than their (R)-antipode, display
favorable pharmacokinetic profiles, exert in vitro vasorelaxant effects on the isolated thoracic aorta,
and exhibit in vivo efficacy in the anesthetized rabbit erectile model.
KEYWORDS: PDE5 inhibitors; Pyrroloquinolones; bioavailability; male erectile dysfunction
(MED).
* Corresponding authors: Hongxiang Lou, Ph.D
Tel: +86-531-88382051; Fax: +86-531-88382019.
E-mail address: louhongxiang@sdu.edu.cn (H. Lou)

1. Introduction
ACCEPTED MANUSCRIPT
Phosphodiesterase 5 (PDE5) is a cGMP-specific PDE that is widely distributed in vascular
smooth muscle cells, lungs, platelets, and corpus cavernosum. PDE5 inhibitors prevent the
hydrolysis of cGMP, maintaining the desired level of cGMP, and are effective for the treatment of
diseases associated with low cGMP levels. Several FDA-approved PDE5 inhibitors, for example,
sildenafil, vardenafil, tadalafil and avanafil (Fig.1), have gained great success for the treatment of
male erectile dysfunction (MED) and pulmonary arterial hypertension (PAH).[1-4] In addition, the
PDE5 inhibitors has also been applied for the treatment of stroke, Raynaud’s disease, and overactive
bladder.[5] Recently, the combined application of PDE5 and HDAC inhibitors was also reported to
be a novel therapeutic and drug design strategy for the treatment of Alzheimer’s disease.[6-8]
Therefore, the seeking for more potent and selective PDE5 inhibitors is still highly desired.
Figure 1. Examples of commercial PDE5 inhibitors.
[Figure 1 here]
The
structures with potent and selective PDE5 inhibitory activity (Fig.2), and TH
structure widely distributed in natural products.[9-11] Several interesting -carboline has been found
in Tribulus terrestris L., a traditional Chinese medicine for the treatment of cardiovascular disease,
and its extracts are claimed to improve sexual performance.[12] Among these -carbolines alkaloids,
perlolyrine (1), [13] accompany with its isomeric structure tribulusterine (2) [14] were prepared by
α
-substituted tetrahyro-
β
-carboline (TH
β
C) related derivatives were proved to be privileged
β
C are also common
β
β

total synthesis and submitted to our preliminary investigation for PDE5 inhibitory activity. Because
ACCEPTED MANUSCRIPT
of the weak PDE5 potency (IC50 = 9.3 and 4.9 µM) and the structural similarity between the
β
-carboline scaffold with the scaffold of tadalafil and other
(Fig. 2[15]), compound 2 was further structural optimized by introducing the characteristic furan
ring into these privileged TH C-related structures. We also referred to the Winterfeldt reaction in the
presence of t-BuOK/O2 for the conversion of PDE5-inhibitor -carbolines (e.g. 3) to the structurally
β-carboline-derived PDE5 inhibitors 3–4
β
β
distinct pyrroloquinolones 5–8,[16, 17] which were identified as potent PDE5 inhibitors (Fig. 2).
Figure 2. Examples of β-carboline and pyrroloquinolone derivatives 1–8.
[Figure 2 here]
In addition to the structure-based pharmaceutical and synthetic analysis of known PDE5
inhibitors, the predictive docking data (Fig. 3, PDB code 1tbf, [18]) also encourage us introducing
the furan fragment of 1–2 to these privileged scaffolds for the discovery of novel chemical entities
as potent and selective PDE5 inhibitors. In accordance with the bioassay results, the predictive
docking data of 1 and the target derivatives are shown in Fig. 3, furan/thiophene rings have higher
electron-rich properties and may lead to better π–π stacking and/or Van der Waals forces with
hydrophobic residues lining the Q2 pocket of PDE5.[19] Previously, most pharmaceutical and
chemical research still focused on the
biological properties of these -furyl/thienyl analogues are valuable for promoting drug discovery
process. To our best knowledge, the PDE5 potency (including structure-activity relationships,
α-phenyl THβCs [10, 20], therefore, systematic study on the
α
druggabilities et al.) of THβCs and pyrroloquinolones with pendant furyl/thienyl have never been
investigated.

ACCEPTED MANUSCRIPT
Figure 3. Crystal structure of PDE5A in complex with Sildenafil (navy) (left, distances are 2.93 Å (left) and 2.84
Å (right)); Predictive binding mode of compounds 1 (navy) and 3 (yellow) (middle, distances NH–O are 2.2 and
1.9 Å, respectively); compounds 7 (navy) and 16a (purple) (right, distances NH–O are both 2.0 Å, respectively)
with PDE5, PDB code 1tbf. [18]
[Figure 3 here]
In present work, bioassay-guided optimization led to the discovery of a series pyrroloquinolone
derivatives with pendant furan/thiophene rings as potent and selective PDE5 inhibitors. The effects
of size, branching, and hydrophobicity of these novel substituents were well studied and concluded
for the SARs. (S)-20a and (S)-20b containing 3-ethylfuran and 3-ethylthiophene substituents,
respectively, were the most potent and selective PDE5 inhibitors. The pharmacokinetic profiles and
in vivo efficacy of (S)-20a and (S)-20b indicated that these compounds are potential drug candidates
for further clinical development.
2. Results and discussion
2.1. Design and Chemistry
The furan fragment in compound 1 was introduced into the literature reported
tetrahydro-β-carboline scaffold and then converted to the related privileged N4-pyrimidine
pyrroloquinolones. Bioassays revealed that the PDE5 inhibitory activity of these pyrroloquinolones
was improved more than 50 fold. Following modification, the furan or thiophene substitution was
carried out according to a biology-oriented synthesis (Fig. 4). Finally, replacement of the
N4-pyrimidine group resulted the optimal compounds.

ACCEPTED MANUSCRIPT
Figure 4. Design and optimization of the PDE5 inhibitors.
[Figure 4 here]
We firstly synthesis compound 1–2 with weak PDE5 inhibitory activity, according to White’s
method (SI, scheme S1).[13] Furthermore, in order to introduce the furan fragment in compound 1–
-
2 into the tetrahydro-β-carboline scaffold, a practical TEMPO oxoammonium salt (T+BF4 )
mediated C–H functionalization of N-pyrimidine/pyridine THβCs was established after detailed
screening (See SI, Table S1). The reaction are clean and efficient, and the reactants could be easily
synthesized according to our previous reports. [21, 22] To seeking potent and selective PDE5
inhibitors, a wide range of structurally distinctive potassium trifluoroborates was prepared and
directly used in the C–H functionalization of N-pyrimidine THβC under optimized condition to
generate diverse -furyl/thienyl TH Cs 10a–l, 11a–u (scheme 1). The nucleophiles mainly consist
α
β
of 3/4/5-substituted alkyl, alkoxyl, alkanol, benzyl and bromide furans or thiophenes, which were
proved to be competent components for the C–H Arylation process. Meanwhile, this condition was
also appropriate for the synthesis of some valuable
TH Cs bearing various electronically R1 moieties or diverse N-pyrimidine/pyridine groups were
also well tolerated in good efficiency (14a–c, 15a–k).
Almost all the novel -furyl/thienyl TH Cs underwent Winterfeldt oxidation process smoothly
α-alkynyl/phenyl THβCs (12a–c, 13a–b).
β
α
β
to provide corresponding pyrroloquinolones 16a–g, 17a–v, 18b–c, 19a–c and 20a–j (Scheme 2)
under the condition of t-BuOK/O2. However, the reaction also completed in the presence of K2O2 (3
equiv.), more dangerous but less yield. The pyrroloquinolone structures were confirmed by X-ray
analysis of lead compound 16e. For the chiral compounds 20a–b, the enantiomeric isomers were
purified using chiral HPLC with a Chiralpak IE column (see SI, Fig. S1). The configuration of

compound these purified enantiomers was determined by optical rotation and circular dichroism
ACCEPTED MANUSCRIPT
spectrum.
Scheme 1. C-H functionalization of N-pyrimidine/pyridine/aroyl THβCs with diverse potassium trifluoroborates.
b
aReaction conditions: THβCs (1 mmol), RBF3K (1.5 mmol), T+BF4− (1.1 mmol), CH2Cl2 (10 mL). Completely
conversion of THβCs was observed after 0.5-1 h. cThese potassium trifluoroborates were used as isomeric mixture
prepared from 3-substituent furan/thiophene, the ratio values represented the proportion of 3-/4-substituted
d
e
products. 10l was obtained after deprotection of 10f. 1.5 mmol T+BF4− was appropriated for phenyl potassium
trifluoroborates (13a–b). fUp to 5-gram scale.
[Scheme 1 here]
Scheme 2. General procedure for Winterfeldt oxidation of α-substituted THβCs; X-ray ORTEP view of compound 16e.a
aReagent and comdition: (a) tBuOK, O2, DMF, r.t. or K2O2, DMF, r.t.
[Scheme 2 here]

2.2. PDE5 Potency and Selectivity for PDE1–4, 6, 11.
ACCEPTED MANUSCRIPT
Subsequently, all derivatives were subjected to biological test for their PDE5 inhibitory activity
according to the literature procedure.[23] Both compound 1 and 2 express only weak PDE5 potency,
maybe because of molecular size is too small to binding to the PDE5 catalytic domain specifically
and firmly. As shown in Table 1-2, most of the C3-substituted pyrroloquinolones exhibited
extremely excellent PDE5 potency at 100 nM, but many α-furyl/thienyl THβCs 10a–l,11a–u,12a–c
(including phenethyl form),14a–c,15a–k showed weak inhibitory activity with IC50 > 100 nM. The
SARs of these furan/thiophene substituents were well concluded.
The natural alkaloid derived alkanol derivatives 16a–b,17a–d (Table 1, entry 1–6) all exhibited
more than 50% inhibition of PDE5 at 100 nM, and the 5-alkanol thiophene analogues 17b–d (Table
1, entry 4–6) also exhibited > 50% inhibition at 10 nM. The PDE5 inhibitory activity of these
alkoxyl derivatives 17e–i (Table 1, entry 7–11) decreased from 84% to 4% inhibition at 100 nM as
the chain lengthened (from 3 to 11 atoms), indicating long-chain derivatives with more than 3 atoms
appeared to reduce the PDE5 activity. Subsequently, the length (0–6 atoms) and substitution position
(C-3/4/5 monosubstituted) of the alkyl chains, as well as the benzyl group, were systematically
studied to provide compounds 16c–g and 17j–u (Table 1, entry 12–28). These results confirmed our
hypothesis and indicated that 5-ethyl furan/thiophene analogues 16e and 17l (Table 1, entry 14 and
19), with 74% and 84% inhibition at 10 nM, respectively, were the best choices for the next
optimization. Also, compound 17v (Table 1, entry 29), containing 5-bromide thiophene strongly
inhibited PDE5, 56% at 10 nM. However, 3-phenylethyl pyrroloquinolones 18a–b were inactive at
100 nM. The IC50 values of compounds 16e and 17l and their N4-modified derivatives 20a–j are
shown in Table 2. Obviously, analogues 20a–e obtained by introducing a pyridine ring exhibited
significantly increased potency against PDE5, and their potencies ranged from 0.5 to 0.8 nM,
regardless of the nitrogen position or the introduction of fluorine. As expected, the (S)-enantiomers

of 20a–b demonstrated stronger inhibitory activity than their antipodes, with a 52- or 3-fold
ACCEPTED MANUSCRIPT
difference, respectively. N4-pyridine and 2-phenylpyrimidine derivatives 20f–j showed weaker
PDE5 inhibition (IC50 = 1.20–6.45 nM), and C7–substituents (Me, OMe and Cl), as in compounds
19a–c (Table 1, entry 30–32), did not increase the PDE5 inhibition.
As shown in Table 2, compounds 16e, 17l, and the most potent compounds 20a–b, accompany
by their enantiomers, were examined for their selectivity for PDE5 relative to other PDE isoenzymes.
All tested compounds had excellent selectivity against PDE1–4 (> 10 000 or > 2 500 fold). Rac-20a
and rac-20b also exhibited good PDE6/5 selectivity (8 and 4 fold), and the selectivity for their
enantiomers was reduced to 1.3–3.9 fold. Finally, compounds (S)-20a and (S)-20b were 5- and
13-fold more selective, respectively, for PDE11.

Table 1. PDE5 potency of 3-furan/thiophene pyrroloquinolone derivatives 16a–g, 17a–v, 19a–c.
ACCEPTED MANUSCRIPT
% PDE5 inhibitiona
X
Entry
No.
R1
R2
100 nM
10 nM
Alkanol and alkoxyl:
1
2
16a
16b
17a
17b
17c
17d
17e
17f
H
H
H
H
H
H
H
H
H
H
H
3-hydroxymethyl
4-hydroxymethyl
3-hydroxymethyl
4-hydroxymethyl
5-hydroxymethyl
5-hydroxyethyl
O
O
S
S
S
S
S
S
S
S
S
57
65
63
87
87
87
84
67
56
18
4
18
21
18
59
65
64
65
–
3
4
5
6
7
5-(methoxymethyl)
5-(2-propoxyethyl)
5-(2-butoxymethyl)
5-(2-(isopentyloxy)ethyl
5-(2-(octyloxy)ethyl
8
9
17g
17h
17i
–
10
11
–
–
3,4,5-Alkyl:
16c
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
5-methyl
5-ethyl
5-pentyl
5-benzyl
H
O
O
O
O
O
S
59
86
94
33
8
–
52
74
–
16d
16e
16f
16g
17j
–
74
92
98
88
75
17
92
81
76
3
–
17k
17l
5-methyl
5-ethyl
5-propyl
5-butyl
S
62
84
60
–
S
17m
17n
17o
17p
17q
17r
17s
17t
S
S
5-hexyl
3-methyl
4-methyl
3,4-dimethyl
3-octyl
S
–
S
66
56
–
S
S
S
–
4-octyl
S
1
–
H (N4-5-
phenylpyrimidin -2-yl)
17u
H
O
61
83
–
Bromide:
29
17v
H
5-Br
S
56
C-7 modification:
30
31
32
19a
19b
19c
Me
5-ethyl
S
S
S
58
21
44
98
38
13
11
74
OMe
Cl
5-ethyl
5-ethyl
Sildenafil
aPDE5 inhibition was determined as the mean of two tests. bNot determined
[Table 1 here]

Table 2. Selectivity data for the N-aryl-modified derivatives 16e, 17l, and 20a–j.
ACCEPTED MANUSCRIPT
IC50a for
PDE5 (nM)
PDE1/
PDE5
PDE2/
PDE5
PDE3/
PDE5
PDE4/
PDE5
PDE6/
PDE5
PDE11/
PDE5
Compd
X
Y
R1
b
16e
17l
O
S
S
S
S
O
O
O
S
S
S
S
S
S
S
S
N
N
N
N
N
N
N
N
N
N
N
C
C
N
N
N
H
H
3.14 ± 0.16
2.50 ± 0.28
0.60 ± 0.01
0.52 ± 0.06
26.9 ± 3.20
0.54 ± 0.06
0.39 ± 0.07
0.97 ± 0.15
0.67 ± 0.06
0.60 ± 0.02
0.82 ± 0.02
2.85 ± 0.35
1.20 ± 0.01
1.25 ± 0.07
6.45 ± 0.49
3.28 ± 0.41
> 10 000
> 10 000
> 10 000
> 2 500
2.0
1.5
8.0
2.5
1.3
4.3
3.9
3.9
–
–
> 10 000
> 10 000
> 10 000
> 2 500
–
c
20a
2-pyridinyl
2-pyridinyl
2-pyridinyl
2-pyridinyl
2-pyridinyl
2-pyridinyl
3-pyridinyl
4-pyridinyl
3-(2-F-pyridinyl)
H
> 10 000
> 10 000
> 10 000
> 2 500
–
(S)-20a
(R)-20a
20b
> 10 000
> 10 000
> 10 000
> 2 500
5.2
–
> 10 000
> 1 500
> 10 000
> 2 500
> 10 000
> 10 000
> 10 000
> 2 500
–
(S)-20b
(R)-20b
20c
> 10 000
> 10 000
> 10 000
> 2 500
13.1
–
> 10 000
> 10 000
> 10 000
> 2 500
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
20d
–
–
20e
–
–
20f
–
–
20g
2-pyridinyl
3,4-OMePh
4-OMePh
Br
–
–
20h
–
–
20i
–
–
20j
–
–
2,3-dihydro
benzofuran
Sildenafilc
7 d
N
2-pyridine
0.40
> 15 000
> 15 000
> 15 000
> 2 500
–
–
3.65 ± 0.57
140
8910
12900
2560
2.0
–
b
c
aIC50 was the mean of three tests. Not determined. For compound 7, literature reported data were listed; for
sildenafil, IC50 values for PDE5–6 were experimental results, selectivity for PDE1–4 were literature data. [24]
[Table 2 here]
2.3. Pharmacokinetic Profiles, in vitro and in vivo Efficacy of (S)-20a and (S)-20b
Pharmacokinetic data of compounds (S)-20a and (S)-20b were obtained using male Wistar rats
dosed at 2 mg/kg (intravenous, iv) or 15 mg/kg (oral, po). Their pharmacokinetic profiles are
presented in Table 3. Compounds (S)-20a and (S)-20b were all orally available and had acceptable
absolute bioavailabilities, F = 24% and 66%, respectively. (S)-20b exhibited improved
bioavailability compared to the C3-phenyl analogues (10-40%[16]). However, the thiophene
derivative (S)-20a had a shorter T1/2 (4.3 h vs. 7.8 h), whereas the furan derivative (S)-20b displayed
a shorter Tmax (1.3 h) and a higher Cmax (46.9 ng/mL). These individual characteristics imply
different clinical applications.

Table 3. Pharmacokinetic profiles of compounds (S)-20a and (S)-20b.
ACCEPTED MANUSCRIPT
Dose
(mg/kg)
AUC0→24h
(µgˑh/L)
AUC0→∞
(µgˑh/L)
Compd
T1/2(h)
Cmax(µg/L)
Tmax (h)
MRT(h)
CL(L/hˑkg)
Vss(L/kg)
F(%)
24.7
2 (iv)
1.9 ± 0.9 307.7 ± 103.5
4.3 ± 1.2 27.5 ± 9.4
2.9 ± 1.5 89.2 ± 19.7
7.8 ± 3.3 46.9 ± 17.6
0.1 ± 0.0 121.1 ± 47.8
2.8 ± 1.3 224.7 ± 126.2
0.1 ± 0.0 67.9 ± 20.4
1.3 ± 0.5 311.8 ± 124.0
125.3 ± 48.7
231.9 ± 131.3
69.3 ± 21.9
2.9 ± 1.4 17.2 ± 6.1
6.7± 1.7 19.0 ± 6.8
72.7 ± 32.5
(S)-20a
15 (po)
2 (iv)
181.3 ± 92.9
124.2 ± 30.5
420.9 ± 120.2
3.1 ± 1.5 32.0 ± 12.4
9.4 ± 2.4 33.0 ± 12.9
(S)-20b
66.3
15 (po)
344.5 ± 149.6
[Table 3 here]
The in vitro vasorelaxant effects of (S)-20a and (S)-20b were evaluated in isolated rabbit
thoracic aorta contracted by norepinephrine. PDE5 inhibitors have vasorelaxant properties that
depend on the endothelium and nitric oxide (NO)-soluble guanylyl cyclase (sGC)-cyclic guanosine
monophosphate pathway. Selective PDE5 inhibition could increase cGMP breakdown and potentiate
the vasorelaxant effect.[25, 26] As shown in Fig. 5, Rac-20a–b and (S)-20a–b exhibited significant
vasorelaxant effects by inducing almost complete aortic relaxation (> 90%) at a concentration of 1
µM, while, their enantiomers (R)-20a–b only produced ~30% response at 10 µM. Considering no
H-bond can be found in the predictive binding mode of (R)-enantiomers and PDE5 (SI, Fig.S326),
we proposed that (R)-20a and (R)-20b cannot bind to PDE5 firmly, and they would be metabolized
or extruded to the extracellular space soon, leading to poor functional response.
Figure 5. Effects of compounds 20a–20b and their enantiomers on norepinephrine (NE)-preconstricted isolated
rabbit aortic rings (compounds Rac-20a, (S)-20a, Rac-20b, and (S)-20b, vehicle, n = 2; (R)-20a and (R)-20b, n =
1).
[Figure 5 here]
Compounds (S)-20a and (S)-20b were then underwent an in vivo study in anesthetized male New
Zealand rabbits model.[27] The intracavernosal pressure (ICP) of electrically stimulated rabbits was
measured after ascending doses of the tested compound were administered via bolus injection. The
ICP should increase after administration of a PDE5 inhibitor in this model as a result of increased

cGMP levels and the subsequent swelling of the penis. The dose-dependent responses are
ACCEPTED MANUSCRIPT
summarized in Fig.6. The ED50 of enantiomers (S)-20a and (S)-20b in vivo were 21.68 µg/kg and
24.21 µg/kg, respectively, and sildenafil was used as the positive control.
Figure 6. In vivo efficacy of (S)-20a and (S)-20b.
[Figure 6 here]
2.4. Molecular Modeling Analysis
Finally, (S)-20a and (S)-20b were docked into the sildenafil-binding site using the GOLD
(Genetic Optimization for Ligand Docking) program (Protein Data Bank (PDB) code 1tbf [18]). The
binding mode of compounds (S)-20a and (S)-20b were further studied through energy minimization.
In the hypothetical protein-ligand complex structure (Fig.7), (S)-20a and (S)-20b exhibited similar
poses and both fit nicely in the active binding site. These compounds were found to form strong
H-bonds with Gln 817, with distances 2.0 Å ((S)-20a) and 2.2 Å ((S)-20b). Moreover, the
quinolinone moieties were able to form π−π stacking interactions with the Phe 820 residues, and the
C3–thiophene/furan substituents could make hydrophobic contacts with the residues lining the
hydrophobic pocket (Q2 pocket) formed by Ala 779, Leu 804, Ile 813 and Met 816. According to the
docking results of enantiomers (Fig.7 and Fig. S326 in SI), we indicated that the different binding
poses of (S)- and (R)-isomers inside the PDE5 binding pocket is the main reason for their distinct
biological activity.

ACCEPTED MANUSCRIPT
Figure 7. Proposed molecular modeling of compounds (S)-20a (navy) and (S)-20b (pink) with PDE5: (Left)
Overall view; (middle) molecular surface of the binding pockets for (S)-20a and (S)-20b with PDE5; and (right)
proposed binding model of (S)-20a and (S)-20b in the sildenafil bonding site; distances were 2.0 and 2.2 Å,
respectively. PDB code 1tbf. [18]
[Figure 7 here]
3. Conclusion
In summary, the discovery of perlolyrine 1, the natural alkaloid with weak PDE5 potency
isolated from the traditional Chinese aphrodisiac plant Tribulus terrestris L., inspired us to diverse
synthesis for seeking more α-furyl THβC derivatives as potent structurally distinctive PDE5
inhibitors. A practical C–H functionalization of N-pyrimidine THβCs was employed for introducing
a wild range of varied furan and thiophene substituents. Next Winterfeldt oxidation leads to a large
amount of C3–furyl/thienyl pyrroloquinolone derivatives 16a–g, 17a–v, 18b–c, 19a–c and 20a–j.
According to subsequent PDE5 assays, SAR analysis and optimization provided
5-ethylfuran/5-ethylthiophene pyrroloquinolones (S)-20a–b as the most potent and selective PDE5
inhibitors. Further experiments demonstrated that (S)-20a and (S)-20b were both orally bioavailable
in preliminary pharmacokinetic studies and exhibited in vitro vasorelaxant activity and in vivo
efficacy in a rabbit model of erectile dysfunction. In addition, the present study also represented that
introducing characteristic furan and thiophene substituents to privileged structures is an effective
strategy for PDE5 inhibitor discovery, and we expect these results would be applied for more drug
discovery process.
4. Experiment section
4.1. General materials and methods.
Melting points were determined with an X-6 melting point apparatus (Beijing TECHInstrument Co. Ltd.) and
are uncorrected. NMR spectra were obtained using a Bruker Avance AVIII-400 spectrometer operating at 400 (1H)

or 101 (13C) MHz. HR-ESI-MS was conducted with an LTQ Orbitrap XL instrument. LC-MS/MS were performed
ACCEPTED MANUSCRIPT
by an API 5500 QTRAP mass spectrometer (Applied Biosystems Sciex, Singapore) with an electrospray ionization
source (ESI) in positive mode. HPLC was performed with an Agilent 1200 system equipped with a G1311A
isopump, a G1322A degasser, and a G1315D DAD detector using an Eclipse XDB-C18 (150 mm × 4.6 mm × 5
µm) and Chiralpak IE (150 mm × 4.6 mm × 5 µm). Column chromatography was conducted on silica gel or
alumina (110–300 mesh). All solvents used were of analytical grade. Precoated silica gel GF254 plates (Qingdao
Haiyang Chemical Co., Ltd.) were used for thin-layer chromatography. All reactions were conducted under a
nitrogen atmosphere with dry solvents under anhydrous conditions. The purities of all biologically evaluated
compounds were established as ≥ 95% by HPLC analysis with UV detection. All the tension and pressure were
measured by the BL-420F Data Acquisition & Analysis System (Chengdu Techman Software Co. Ltd.) equipped
with corresponding transducer.
4.2. General synthetic procedures.
The general procedures for the synthesis of diverse α-furyl/thienyl tetrahyro-β-carbolines and corresponding
pyrroloquinolones analogues were also descripted with typical examples. Compounds 10a–l, 11a–u, 12a–c, 13a–b
were synthesized from starting material 9g and variousborane derivatives S10a–l, S11a–u, S12a–c, S14a–b (SI,
Scheme S2); Compounds 14a–c, 15a–c, 15e–g, 15i–k were synthesized from potassium
(5-ethylthiophene-2-yl)trifluoroborate and 9l,o,k,f–g,n,h–j (SI, Scheme S3); Compounds 15d was synthesized
from potassium (furan-2-yl)trifluoroborate and 9e (SI, Scheme S3); Compounds 15h was synthesized from
potassium (5-ethylfuran-2-yl)trifluoroborate and 9n (SI, Scheme S3). Pyrroloquinolones 16a–g were synthesized
from 10f–g,a–e; compounds 17a–v were synthesized from 11h–i,n,q,o,r,p,t,s,a–f,j–k,g,l–m, 15d, 11u; compounds
18b–c and 19a–c were synthesized from 12b–c (Phenethyl form), 14a–c, compounds 20a–j were synthesized from
15g–h,i–k,a–b,f,e,c. The analytical data, HPLC assays, X-ray crystallographic analysis for individual derivatives
can be found in the supporting information.
1-(5-ethylthiophen-2-yl)-2-(5-(pyridin-2-yl)pyrimidin-2-yl)-2,3,4,9-tetrahydro-1H-pyrido[3,4-b]indole (15g).
2-(5-(pyridin-2-yl)pyrimidin-2-yl)-2,3,4,9-tetrahydro-1H-pyrido[3,4-b]indole 9g (327 mg, 1 mmol, 1 eq) was
dissolved in CH2Cl2, (327 mg, 1.5 mmol, 1.5 eq). Then potassium (5-ethylthiophene-2-yl)trifluoroborate (327 mg,
1.5 mmol, 1.5 eq) and 2,2,6,6-tetramethyl-piperidine N-oxide salt (T+BF4−) (153 mg, 1 mmol, 1 eq) were added in
sequence. The mixture was allowed to be stirred for 30 min at room temperature and quenched by adding saturated
NaHCO3 solution at 0 °C. The organic phase was separated and the aqueous layer was extracted with
dichloromethane (2 × 15 mL). The combined organic phase was dried, concentrated and chromatographed on silica
to give compound 15i (340 mg, 78%) as white solid. 1H NMR (400 MHz, CDCl3) δ 8.89 (s, 2H), 8.52 (d, J = 4.7
Hz, 1H), 8.10 (s, 1H), 7.60 (td, J = 7.8, 1.6 Hz, 1H), 7.46 (d, J = 6.9 Hz, 2H), 7.24 – 7.12 (m, 2H), 7.12 – 6.98 (m,
3H), 6.67 (d, J = 3.4 Hz, 1H), 6.45 (d, J = 3.2 Hz, 1H), 5.05 (dd, J = 13.3, 4.7 Hz, 1H), 3.39 (td, J = 13.2, 4.2 Hz,

13C
1H), 3.07 – 2.85 (m, 1H), 2.80 (dd, J A= C15C.3E, 3P.4THEzD, 1MH),A2N.63U(Sq,CJR=I7P.T5 Hz, 2H), 1.13 (t, J = 7.5 Hz, 3H).
NMR (101 MHz, CDCl3) δ 160.8, 156.5, 153.6, 149.9, 148.1, 140.1, 136.9, 136.3, 132.4, 126.9, 126.8, 122.4,
122.1, 121.8, 121.6, 119.5, 118.8, 118.5, 111.1, 110.7, 50.0, 37.7, 23.5, 21.4, 15.9. ESI-MS m/z 438 [M + H]+,
HR-EI-MS: m/z 438.1751 [M + H]+, (calcd for C26H24N5S+, 438.1747).
3-(5-ethylthiophen-2-yl)-2-(5-(pyridin-2-yl)pyrimidin-2-yl)-1,2,3,4-tetrahydro-9H-pyrrolo[3,4-b]quinolin-9-on
e (20a).
Compound 15g (178 mg, 0.4 mmol, 1 eq) was dissolved in 8 mL DMF and added t-BuOK (136 mg, 0.3 mmol,
3 eq). The reaction was bubbled with oxygen and stirred for 8 h before water (10 mL) was added. Then the mixture
was extracted with EtOAc (5 × 30 mL), and the organic layer was combined, dried (Na2SO4), filtered, and
1
concentrated to yield compound 20a as light yellow solid (38 mg, 21%), mp > 250 °C. H NMR (400 MHz,
DMSO) δ 12.19 (s, 1H), 9.12 (s, 2H), 8.62 (d, J = 4.1 Hz, 1H), 8.17 (d, J = 8.0 Hz, 1H), 7.94 (d, J = 7.9 Hz, 1H),
7.85 (t, J = 6.9 Hz, 1H), 7.64 (dt, J = 14.5, 7.9 Hz, 2H), 7.41 – 7.30 (m, 2H), 7.12 (d, J = 3.2 Hz, 1H), 6.71 (s, 2H),
4.88 (d, J = 14.1 Hz, 1H), 4.73 (dd, J = 13.8, 2.2 Hz, 1H), 2.69 (q, J = 7.4 Hz, 2H), 1.13 (t, J = 7.5 Hz, 3H). 13C
NMR (101 MHz, DMSO) δ 173.6, 159.5, 156.9, 153.3, 153.0, 150.1, 150.0, 147.3, 141.0, 140.4, 137.7, 132.2,
127.1, 125.6, 125.3, 123.7, 122.7, 122.3, 119.5, 119.1, 118.9, 114.0, 61.8, 50.2, 23.2, 16.0. ESI-MS m/z 452 [M +
H]+. ESI-MS m/z 452 [M + H]+, HR-EI-MS: m/z 452.1542 [M + H]+, (calcd for C26H22N5OS+, 452.1540). The
enantiomers are separated by the chiral HPLC, and their NMR and MS data were the same as reported above.
(S)-(-)-20a. Light yellow needle crystal, mp > 250 °C, tR = 4.423 min. [α]20D = -244.31, c = 1.02 in methanol; CD
(MeOH) λmax (∆ε) 295 (-0.82), 272 (-1.60), 252 (+2.49), 230 (-1.76) nm. (R)-(+)-20a. Light yellow crystal, mp >
250 °C, tR = 4.970 min. [α]20D = +213.27, c = 1.15 in methanol; CD (MeOH) λmax (∆ε) 296 (-0.14), 271 (+0.98),
253 (-2.10), 227 (+1.42) nm.
3-(5-ethylfuran-2-yl)-2-(5-(pyridin-2-yl)pyrimidin-2-yl)-1,2,3,4-tetrahydro-9H-pyrrolo[3,4-b]quinolin-9-one
(20b).
Compound 20b was synthesized according to the general procedure from 9g with potassium
(5-ethylfuran-2-yl)trifluoroborate. Yield 28%. White solid, mp > 250 °C. 1H NMR (400 MHz, DMSO) δ 12.17 (s,
1H), 9.11 (s, 2H), 8.63 (d, J = 4.3 Hz, 1H), 8.17 (d, J = 8.0 Hz, 1H), 7.94 (d, J = 7.9 Hz, 1H), 7.85 (t, J = 7.5 Hz,
1H), 7.75 – 7.56 (m, 2H), 7.34 (dt, J = 12.3, 6.7 Hz, 2H), 6.53 (s, 1H), 6.48 (d, J = 1.8 Hz, 1H), 6.06 (s, 1H), 4.87
(d, J = 13.9 Hz, 1H), 4.76 (dd, J = 13.9, 2.2 Hz, 1H), 1.06 (t, J = 7.5 Hz, 3H). 13C NMR (101 MHz, DMSO) δ
173.5, 159.7, 157.6, 156.9, 153.0, 150.1, 149.7, 148.4, 141.0, 137.7, 132.1, 125.7, 125.2, 123.7, 122.7, 122.2,
119.4, 119.1, 114.6, 110.3, 105.6, 60.2, 50.4, 21.2, 12.1. ESI-MS m/z 436 [M + H]+, HR-EI-MS: m/z 436.1765 [M
+
+ H]+, (calcd for C26H22N5O2 , 436.1768). (S)-(-)-20b. Yellow crystal, mp > 250 °C, tR = 4.105 min. [α]20
=
D
-217.15, c = 1.20 in methanol; CD (MeOH) λmax (∆ε) 298 (-0.99), 269 (-1.73), 240 (+2.87), 226 (-0.84), 209
(-3.54) nm. (R)-(+)-20b. Light yellow crystal, mp > 250 °C, tR = 4.509 min. [α]20D = +190.13, c = 1.03 in methanol;

CD (MeOH) λmax (∆ε) 299 (+1.15), 2A69C(+C1E.8P9)T, 2E4D0 (-M2.8A0)N, 2U26S(C+1R.0I2P)T, 209 (+4.00) nm.
4.3. Bioassay.
4.3..1 PDE assays.
These assays were carried out as described previously.[23] All of the enzymatic reactions were conducted at
25°C for 60 min. The 50 µL reaction mixture contains 40 mM MOPS at pH 7.5, 0.5 mM EDTA, 15 mM MgCl2,
0.15 mg/mL BSA, 1 mM DTT, 0.05% ProClin 200, 15 ng/mL PDE5 (or other PDE isoforms) and 100 nM
FAM-cyclic-3’,5-GMP (or FAM-cyclic-3’,5’-AMP for different PDE isoforms). The compounds were diluted in 10%
DMSO, and 5 µL of the dilution was added to a 50 µL reaction for a final concentration of 1% DMSO in all
reactions. The reaction mixture was incubated at 25 °C for 1 h. Then, 100 µL of diluted binding agent was added to
each well and incubated at 25 °C for 1 h with slow shaking. The fluorescence polarization of the sample was read
using an excitation filter of 360 nm and an emission filter of 480 nm. The IC50 values were calculated using
nonlinear regression with a normalized dose-response fit using Prism 6 (GraphPad software, San Diego, CA). The
PDEs employed included PDE1A (BPS Bioscience Inc. #60010), PDE2A (BPS Bioscience Inc. #60021), PDE3B
(BPS Bioscience Inc. #60031), PDE4A1A (BPS Bioscience Inc. #60040), PDE5A (BPS Bioscience Inc. #60050),
PDE6C (BPS Bioscience Inc. #60060), and PDE11A (BPS Bioscience Inc. #60110).
4.3.2. Pharmacokinetic study.
All the animal procedures in the experimental protocol were reviewed and approved by the Institutional
Animal Care Committee of Shandong University. Pharmacokinetic analysis of (S)-20a and (S)-20b was conducted
on male Wistar rats (body weight of 228−249 g). For each tested compound, the rats were separated into two
groups and fasted for 12 h before po or iv administration, for which 0.5% methylcellulose and 10% HPBCD were
used as vehicles, respectively. Blood samples were collected at 0, 0.5, 1, 1.5, 2, 4, 6, 12 and 24 h after po
application or 0, 0.1, 0.15, 0.5, 0.75, 1,1.5, 2, 4, 6, 12 and 24 h after iv application. After centrifugation at 14000
rpm and 4 °C, the plasma samples were frozen at −80 °C. Quantification in the rat plasma was determined using an
LC–MS/MS method described previously[28], and the pharmacokinetic parameters were calculated using Drug
and Statistics Software version 2.0 (DAS 2.0, Mathematical Pharmacology Professional Committee of China,
Shanghai, China).
4.3.3 Vasorelaxant effect on norepinephrine-pretreated isolated rabbit thoracic aorta.
The tension and pressure were measured using a BL-420F Data Acquisition & Analysis System (Chengdu
Techman Software Co. Ltd.). New Zealand white rabbits were stunned, and their aorta were carefully isolated, cut,
and incubated as described in the literature.[25, 26] The aorta rings were then perfused in Krebs buffer, stretched
and maintained at a tension of 2 g (0%), and the tension was allowed to equilibrate for 2 h. The aortic rings was
first contracted with norepinephrine (1 µM) to the maximal response (100%), and the tested compounds were
added carefully at 15 min intervals. The ring tensions were recorded at 15 min after each dosing. The vasorelaxant

effect was determined as a percentaAgeCoCf EthPeTreEsDponMseAcaNlcUulSatCedRuIsPinTg the following formula: Response =
(Maximal tension (g) - Record tension (g)) / (Maximal tension (g) - 2 g) × 100%.
4.3.4. In vivo efficacy study.
The anesthetized rabbit model was established according to the literature.[27] Male New Zealand white rabbits
with body weights of 1.9-2.2 kg were fasted for 12 h and anesthetized using sodium pentobarbital. The penis was
denuded of skin down to the base, and a 22-gauge scalp needle, attached by flexible catheter to a pressure
transducer, was inserted into the corpus cavernosum to record the ICP changes in response to pelvic nerve
stimulation. The stimulation output set was at 5 V/10 V with 1 ms pulse duration, 2 Hz frequency and 1 min
stimulation. The arterial blood pressure was recorded by the artery intubation, and the mean arterial pressure (MAP)
was determined. The area under curve (AUC) for the ICP at each stimulation was computed, and the ICP/MAP
was calculated as the response for erectile function. The baseline (0%) was established before dosing with the
tested compound, and 100 mg/kg sildenafil, which was shown to induce maximal ICP/MAP and was established as
100% response for each rabbit, was injected intravenously at the end of each experiment. The tested compounds
were administered via bolus injection in ascending dosages.
4.3.5. Docking.
Docking studies were performed using SYBYL-X and the GOLD program as described [29], employing the
crystal structure of PDE5 in a complex with sildenafil downloaded from the PDB (PDB code 1tbf).
Acknowledgement
This work was financially supported by the National Natural Science Foundation of China (81630093 and
81473107), the Science and Technology Major Project of Shandong Province (No. 2015ZDJS04001) and Focus on
Research And Development Plan In Shandong Province (2017CXGC1403).
Reference and notes
[1] D.J. Dale, P.J. Dunn, C. Golightly, M.L. Hughes, P.C. Levett, A.K. Pearce, P.M. Searle, W. Gordon, A.S. Wood, The
chemical development of the commercial route to sildenafil:ꢀ A case history, Org. Process Res. Dev., 4 (2000) 17–22.
[2] D.H. Maurice, H. Ke, F. Ahmad, Y. Wang, J. Chung, V.C. Manganiello, Advances in targeting cyclic nucleotide
phosphodiesterases, Nat. Rev. Drug Discovery, 13 (2014) 290–314.
[3] J. Liu, A. Maisonial-Besset, B. Wenzel, D. Canitrot, A. Baufond, J.M. Chezal, P. Brust, E. Moreau, Synthesis and
in vitro evaluation of new fluorinated quinoline derivatives with high affinity for PDE5: Towards the development of
new PET neuroimaging probes, Eur. J. Med. Chem., 136 (2017) 548–560.
[4] J. Fiorito, F. Saeed, H. Zhang, A. Staniszewski, Y. Feng, Y.I. Francis, S. Rao, D.M. Thakkar, S.X. Deng, D.W. Landry,

Synthesis of quinoline derivatives: discoAveCryCofEaPpTotEenDt anMd AselNecUtivSe CphRosIpPhoTdiesterase 5 inhibitor for the treatment of
Alzheimer's disease, Eur. J. Med. Chem., 60 (2013) 285–294.
[5] P. Sandner, J. Hütter, H. Tinel, K. Ziegelbauer, E. Bischoff, PDE5 inhibitors beyond erectile dysfunction, Int. J.
Impotence Res., 19 (2007) 533–543.
[6] O. Rabal, J.A. Sánchez-Arias, M. Cuadrado-Tejedor, I. De Miguel, M. Pérez-González, C. García-Barroso, A. Ugarte,
A.E. De Mendoza, E. Sáez, M. Espelosin, U. Susana, T. Haizhong, W. Wei, X. Musheng, A. Garcia-Osta, J. Oyarzabal,
Design, synthesis, and biological evaluation of first-in-class dual acting histone deacetylases (HDACs) and
phosphodiesterase 5 (PDE5) inhibitors for the treatment of Alzheimer's disease, J. Med. Chem., 59 (2016) 8967–9004.
[7] J.A. Sánchez-Arias, O. Rabal, M. Cuadrado-Ttejedor, I. De Miguel, M. Pérez-González, A. Ugarte, E. Sáez, M.
Espelosin, S. Ursua, T. Haizhong, W. Wei, X. Musheng, A. Garcia-Osta, J. Oyarzabal, Impact of scaffold exploration on
novel dual-acting histone deacetylases (HDACs) and phosphodiesterase 5 (PDE5) inhibitors for the treatment of
Alzheimer's Disease, ACS Chem. Neurosci., 8 (2017) 638–661.
[8] J. Fiorito, J. Vendome, F. Saeed, A. Staniszewski, H. Zhang, S. Yan, S.X. Deng, O. Arancio, D.W. Landry,
Identification of a Novel 1,2,3,4-tetrahydrobenzo[b][1,6]naphthyridine Analog as a Potent Phosphodiesterase 5 Inhibitor
with Improved Aqueous Solubility for the Treatment of Alzheimer's Disease, J. Med. Chem., 60 (2017) 8858–8875.
[9] N.S. Ahmed, B.D. Gary, G.A. Piazza, H.N. Tinsley, S. Laufer, A.H. Abadi, A novel access to arylated and
heteroarylated β-carboline based PDE5 inhibitors, Med. Chem., 6 (2010) 374–387.
[10] N.S. Ahmed, A.H. Ali, S.M. El-Nashar, B.D. Gary, A.M. Fajardo, H.N. Tinsley, G.A. Piazza, M. Negri, A.H. Abadi,
Exploring the PDE5 H-pocket by ensemble docking and structure-based design and synthesis of novel β-carboline
derivatives, Eur. J. Med. Chem., 57 (2012) 329–343.
[11] H.A. Mohamed, N.M.R. Girgis, R. Wilcken, M.R. Bauer, H.N. Tinsley, B.D. Gary, G.A. Piazza, F.M. Boeckler, A.H.
Abadi, Synthesis and molecular modeling of novel tetrahydro-β-carboline derivatives with phosphodiesterase 5
inhibitory and anticancer properties, J. Med. Chem., 54 (2011) 495–509.
[12] M. Dell’Agli, G.V. Galli, E.D. Cero, F. Belluti, R. Matera, E. Zironi, G. Pagliuca, E. Bosisio, Potent inhibition of
human phosphodiesterase-5 by icariin derivatives, J. Nat. Prod., 71 (2008) 1513–1517.
[13] J.B. Bremner, W. Sengpracha, I. Southwell, C. Bourke, B.W. Skelton, A.H. White, A revised structure for the
alkaloid, tribulusterine, from Tribulus Terrestris L., Aus. J. Chem., 57 (2004) 273–276.
[14] T.S. Wu, L.S. Shi, S.C. Kuo, Alkaloids and other constituents from Tribulus terrestris, Photochemistry, 50 (1999)
1411–1415.
[15] Z. Sui, J. Guan, M.J. Macielag, W. Jiang, Y. Qiu, P. Kraft, S. Bhattacharjee, T.M. John, E. Craig, D.
Haynes-Johnson, J. Clancy, Synthesis and biological activities of novel β-Carbolines as PDE5 inhibitors, Bioorg. Med.
Chem. Lett., 13 (2003) 761–765.
[16] W. Jiang, J. Guan, M.J. Macielag, S. Zhang, Y. Qiu, P. Kraft, S. Bhattacharjee, T.M. John, D. Haynes-Johnson, L.

Scott, Z. Sui, Pyrroloquinolone PDE5 inhibitors with improved pharmaceutical profiles for clinical studies on erectile
ACCEPTED MANUSCRIPT
dysfunction, J. Med. Chem., 48 (2005) 2126–2133.
[17] Z. Sui, J. Guan, M.J. Macielag, W. Jiang, S. Zhang, Y. Qiu, P. Kraft, S. Bhattacharjee, T.M. John, D.
Haynes-Johnson, Pyrimidinylpyrroloquinolones as highly potent and selective PDE5 inhibitors for treatment of erectile
dysfunction, J. Med. Chem., 45 (2002) 4094–4096.
[18] K.Y. Zhang, G.L. Card, Y. Suzuki, D.R. Artis, D. Fong, S. Gillette, D. Hsieh, J. Neiman, B.L. West, C. Zhang, A
glutamine switch mechanism for nucleotide selectivity by phosphodiesterases, Mol. Cell, 15 (2004) 279–286.
[19] D.S. El-Gamil, N.S. Ahmed, B.D. Gary, G.A. Piazza, M. Engel, R.W. Hartmann, A.H. Abadi, Design of novel
β-carboline derivatives with pendant 5-bromothienyl and their evaluation as phosphodiesterase-5 inhibitors, Arch.
Pharm., 346 (2013) 23–33.
[20] A.H. Abadi, B.D. Gary, H.N. Tinsley, G.A. Piazza, M. Abdelhalim, Synthesis, molecular modeling and biological
evaluation of novel tadalafil analogues as phosphodiesterase 5 and colon tumor cell growth inhibitors, new
stereochemical perspective, Cheminform, 45 (2010) 1278.
[21] H. Zheng, Y. Dong, L. Li, B. Sun, L. Liu, H. Yuan, H. Lou, Novel benzo[a]quinolizidine analogs induce cancer cell
death through paraptosis and apoptosis, J. Med. Chem., 59 (2016) 5063–5067.
[22] Z. Xie, L. Liu, W. Chen, H. Zheng, Q. Xu, H. Yuan, H. Lou, Practical metal-free C(sp3)-H functionalization:
Construction of structurally diverse α-substitutedn-benzyl andn-allyl carbamates, Angew. Chem. In. Ed., 53 (2014)
3904–3908.
[23] W. Huang, Y. Zhang, J.R. Sportsman, A fluorescence polarization assay for cyclic nucleotide phosphodiesterases, J.
Biomol. Screen., 7 (2002) 215–222.
[24] J.C. Lanter, Z. Sui, M.J. Macielag, J.J. Fiordeliso, W. Jiang, Y. Qiu, S. Bhattacharjee, P. Kraft, T.M. John, D.
Haynes-Johnson, E. Craig, J. Clancy, Structure-activity relationships of N-acyl pyrroloquinolone PDE-5 inhibitors, J.
Med. Chem., 47 (2004) 656–662.
[25] M. Mojzych, M. Kubacka, S. Mogilski, B. Filipek, E. Fornal, Relaxant effects of selected sildenafil analogues in the
rat aorta, J. Enzym. Inhib. Med. Ch., 31 (2016) 1–8.
[26] J.B. Salom, M.C. Burguete, F.J. Pérez-Asensio, M. Castelló-Ruiz, G. Torregrosa, E. Alborch, Relaxant effect of
sildenafil in the rabbit basilar artery, Vas. Pharmacol., 44 (2006) 10–16.
[27] S. Choi, L. O'Connell, K. Min, N.N. Kim, R. Munarriz, I. Goldstein, E. Bischoff, A.M. Traish, Efficacy of
vardenafil and sildenafil in facilitating penile erection in an animal model, J. Androl., 23 (2002) 332–337.
[28] Y. Gao, Q. Sun, D. Liu, B. Ma, H. Zhao, Z. Fang, H. Wang, H. Lou, A sensitive LC–MS/MS method to quantify
methylergonovine in human plasma and its application to a pharmacokinetic study, J. Chromatogr. B, 1011 (2016) 62–
68.
[29] B. Sun, L. Li, Q.W. Hu, F. Xie, H.B. Zheng, H.M. Niu, H.Q. Yuan, H.X. Lou, Design, synthesis and biological

evaluation of novel macrocyclic bisbibenzyl analogues as tubulin polymerization inhibitors, Eur. J. Med. Chem., 121
ACCEPTED MANUSCRIPT
(2016) 484–499.

List of captions:
Figure 1. Examples of commercial PDEA5 iCnhCibEitoPrsT. ED MANUSCRIPT
Figure 2. Examples of β-carboline and pyrroloquinolone derivatives 1–8.
Figure 3. Crystal structure of PDE5A in complex with Sildenafil (navy) (left, distances are 2.93 Å (left) and 2.84 Å
(right)); Predictive binding mode of compounds 1 (navy) and 3 (yellow) (middle, distances NH–O are 2.2 and 1.9 Å,
respectively); compounds 7 (navy) and 16a (purple) (right, distances NH–O are both 2.0 Å, respectively) with PDE5,
PDB code 1tbf. [18]
Figure 4. Design and optimization of the PDE5 inhibitors.
Scheme 1. C–H functionalization of N-pyrimidine/pyridine/aroyl THβCs with diverse potassium trifluoroborates.
b
aReaction conditions: THβCs (1 mmol), RBF3K (1.5 mmol), T+BF4− (1.1 mmol), CH2Cl2 (10 mL). Completely
c
conversion of THβCs was observed after 0.5-1 h. These potassium trifluoroborates were used as isomeric mixture
prepared from 3-substituent furan/thiophene, the ratio values represented the proportion of 3-/4-substituted products.
e
d10l was obtained after deprotection of 10f. 1.5 mmol T+BF4− was appropriated for phenyl potassium trifluoroborates
(13a–b). fUp to 5-gram scale.
Scheme 2. General procedure for Winterfeldt oxidation of α-substituted THβCs; X-ray ORTEP view of compound 16e.a
aReagent and comdition: (a) tBuOK, O2, DMF, r.t. or K2O2, DMF, r.t.
Table 1. PDE5 potency of 3-furan/thiophene pyrroloquinolone derivatives 16a–g, 17a–v, 19a–c.
aPDE5 inhibition was determined as the mean of two tests. bNot determined
Table 2. Selectivity data for the N-aryl-modified derivatives 16e, 17l, and 20a–j.
b
c
aIC50 was the mean of three tests. Not determined. For compound 7, literature reported data were listed; for sildenafil,
IC50 values for PDE5–6 were experimental results, selectivity for PDE1–4 were literature data. [24]
Table 3. Pharmacokinetic profiles of compounds (S)-20a and (S)-20b.

Figure 5. Effects of compounds 20a–20b and their enantiomers on norepinephrine (NE)-preconstricted isolated rabbit
ACCEPTED MANUSCRIPT
aortic rings (compounds Rac-20a, (S)-20a, Rac-20b, and (S)-20b, vehicle, n = 2; (R)-20a and (R)-20b, n = 1).
Figure 6. In vivo efficacy of (S)-20a and (S)-20b.
Figure 7. Proposed molecular modeling of compounds (S)-20a (navy) and (S)-20b (pink) with PDE5: (Left) Overall
view; (middle) molecular surface of the binding pockets for (S)-20a and (S)-20b with PDE5; and (right) proposed
binding model of (S)-20a and (S)-20b in the sildenafil bonding site; distances were 2.0 and 2.2 Å, respectively. PDB
code 1tbf. [18]

ACCEPTED MANUSCRIPT
Highlights
ò
ò
ò
ò
Diversified synthesis of pyrroloquinolones based on natural alkaloid perlolyrine 1.
Practical C–H functionalization for the synthesis of diverse α-furyl/thienyl THβCs.
Systematic SAR investigation of furyl/thienyl pyrroloquinolones for PDE5.
The candidates (S)-20a–b were the most potent with good pharmacokinetic profiles.