Natural phosphodiesterase 5 (PDE5) inhibitors: a computational approach

Abstract

In 1998, sildenafil was marketed as the first FDA-approved oral drug for the treatment of erectile dysfunction (ED). During the last two decades, the commercialization of other synthetic phosphodiesterase 5 (PDE5) inhibitors has been paralleled by the rise of remedies based on natural molecules from different chemical classes (flavonoids, polyphenols and alkaloids in general). In this work, a set of in silico tools were applied to study a panel of 30 natural compounds claimed to be effective against ED in the scientific literature or in folk medicine. First, pharmacokinetic properties were analysed to exclude the compounds lacking in specific drug-like features. Estimated binding energy for PDE5 and selectivity towards other PDE isoforms were then considered to highlight some promising molecules. Finally, a detailed structural investigation of the interaction pattern with PDE in comparison with sildenafil was conducted for the best performing compound of the set.

Full text

Full Terms & Conditions of access and use can be found at
https://www.tandfonline.com/action/journalInformation?journalCode=gnpl20
Natural Product Research
Formerly Natural Product Letters
ISSN: 1478-6419 (Print) 1478-6427 (Online) Journal homepage: https://www.tandfonline.com/loi/gnpl20
Natural phosphodiesterase 5 (PDE5) inhibitors: a
computational approach
Alberto Ongaro, Giuseppe Zagotto, Maurizio Memo, Alessandra Gianoncelli
& Giovanni Ribaudo
To cite this article: Alberto Ongaro, Giuseppe Zagotto, Maurizio Memo, Alessandra Gianoncelli
& Giovanni Ribaudo (2019): Natural phosphodiesterase 5 (PDE5) inhibitors: a computational
approach, Natural Product Research, DOI: 10.1080/14786419.2019.1619726
To link to this article: https://doi.org/10.1080/14786419.2019.1619726
View supplementary material
Published online: 29 May 2019.
Submit your article to this journal
Article views: 35
View Crossmark data

SHORT COMMUNICATION
Natural phosphodiesterase 5 (PDE5) inhibitors:
a computational approach
Alberto Ongaro
a
, Giuseppe Zagotto
b
, Maurizio Memo
a
, Alessandra Gianoncelli
a
and Giovanni Ribaudo
b
a
Department of Molecular and Translational Medicine, University of Brescia, Brescia, Italy;
b
Department of Pharmaceutical and Pharmacological Sciences, University of Padova, Padova, Italy
ABSTRACT
In 1998, sildenafil was marketed as the first FDA-approved oral
drug for the treatment of erectile dysfunction (ED). During the
last two decades, the commercialization of other synthetic
phosphodiesterase 5 (PDE5) inhibitors has been paralleled by the
rise of remedies based on natural molecules from different chem-
ical classes (flavonoids, polyphenols and alkaloids in general). In
this work, a set of in silico tools were applied to study a panel of
30 natural compounds claimed to be effective against ED in the
scientific literature or in folk medicine. First, pharmacokinetic
properties were analysed to exclude the compounds lacking in
specific drug-like features. Estimated binding energy for PDE5 and
selectivity towards other PDE isoforms were then considered to
highlight some promising molecules. Finally, a detailed structural
investigation of the interaction pattern with PDE in comparison
with sildenafil was conducted for the best performing compound
of the set.
ARTICLE HISTORY
Received 15 March 2019
Accepted 12 May 2019
KEYWORDS
PDE5; erectile dysfunction;
flavonoids; sildena-
fil; docking
CONTACT Giovanni Ribaudo giovanni.ribaudo@unipd.it Department of Pharmaceutical and Pharmacological
Sciences, University of Padova, Via Marzolo 5, 35131 Padova, Italy
Supplemental data for this article can be accessed at https://doi.org/10.1080/14786419.2019.1619726.
ß2019 Informa UK Limited, trading as Taylor & Francis Group
NATURAL PRODUCT RESEARCH
https://doi.org/10.1080/14786419.2019.1619726

1. Introduction
Sildenafil was the first oral drug acting through the inhibition of phosphodiesterase 5
(PDE5) approved by Food and Drug Administration (FDA) for the treatment of erectile
dysfunction (ED) (Goldstein et al. 1998; Goldstein 2000; Yafi et al. 2018). PDE5 inhibi-
tors contrast the degradation of cGMP promoting the relaxation of penile smooth
muscle and are currently ranked in terms of potency and of selectivity with respect to
other PDEs, which differ in substrate specificity, kinetic properties and tissue distribu-
tion (Bischoff 2004; Ribaudo et al. 2016). Flushing, headache, dyspepsia back pain,
myalgia hearing loss, visual disturbances, priapism and, in some specific cases, melan-
oma emerged to be the most common adverse effects related to the use of PDE5
inhibitors, which to some extent are due to interference with other PDE isoforms
(especially 1, 3, 6 and 11, among 11 known families) and other molecular mechanisms
(Bischoff 2004; Yafi et al. 2018). The fruitful development of PDE5 inhibitors during the
last two decades was paralleled by the rise of alternative treatments based on natural
remedies, which are often perceived as safer.
2. Results and discussion
Several natural compounds belonging to different chemical classes were reported to
show activity against ED and, although other molecular mechanisms may be involved,
inhibitory properties towards PDE5 have been claimed (Pavan et al. 2015; Abusnina
and Lugnier 2017).
Previous in silico studies investigated the interaction of single classes of natural
molecules, mainly polyphenols, with PDE5 (Srivani et al. 2007; Chen 2009; Ribaudo
et al. 2017; Arya et al. 2018). Moreover, concerning computer-aided drug discovery
aimed at highlighting novel synthetic PDE5 inhibitors, T€
om€
ori et al. presented a com-
prehensive in silico approach, including screening physico-chemical parameters, 2D
similarity and 3D docking (T€
om€
ori et al. 2012).
In the present work, 8 PDE isoforms and a selection of 30 natural compounds from
different plants and belonging to various chemical classes, which were reported as
effective in vitro and in vivo against ED, were considered. Remedies from traditional
medicine and phytocomplexes that were recently re-studied from a more rational
point of view were also included. Among flavonoids, icariin, apigenin, luteolin, naringe-
nin, quercetin, osajin, pomiferin, scandenone, auriculasin, and kraussianones 2-5 were
selected (Pavan et al. 2015; Ribaudo et al. 2015; Abusnina and Lugnier 2017; Oboh
et al. 2017). Some polyphenols and, more specifically, organic acids, namely forskolin,
resveratrol, gallic acid, chlorogenic acid, caffeic acid, ellagic acid, xanthohumol, and
EGCG were chosen and introduced in the study (Pavan et al. 2015; Oboh et al. 2017).
In addition to these two major classes, other alkaloids and anthocyanidins, such as del-
phinidin, maca 2, osthol, ginsenoside Re, caffeine, yohimbine, berberine, macaridine
and tyramine were considered (West and Krychman 2015; Abusnina and Lugnier 2017;
Ribaudo et al. 2018)(Figure 1). Sildenafil was selected as the reference molecule
throughout the in silico routine.
First, the predicted pharmacokinetic properties of the natural molecules were calcu-
lated using Molinspiration Cheminformatics software (https://www.molinspiration.com).
2 A. ONGARO ET AL.

According to the guidelines from the most recent reports on drug-likeness, com-
pounds with TPSA >140 Å
2
and/or MW >800 Da (chlorogenic acid, EGCG, ellagic
acid, ginsenoside Re and icariin) were not admitted to the second step of the study,
since their expected cellular permeability would be very limited (Petit et al. 2012;
Matsson and Kihlberg 2017). Description of the computed parameters, detailed results
and selection criteria are reported in the Supplementary Material (Table S1).
Concerning the docking studies, the PDE isoforms available in the RCSB Protein
Data Bank (www.rcsb.org) were considered, in order to provide a preliminary selectiv-
ity profile towards PDE5. Besides PDE5 (2H44), compounds were docked to PDE1
(5W6E), PDE2 (4JIB), PDE4 (1XOS), PDE6 (3JWQ), PDE7 (4PM0), PDE9 (3JSW) and PDE10
(5K9R) following the protocol described in the Supplementary Material. Results are
reported in Table 1.
Two major criteria were adopted to identify the most promising compounds. First,
taking the estimated DG value for sildenafil towards PDE5 as reference (10.1 kcal/
mol), the molecules showing better binding energy values were highlighted. Second,
the molecules with a selectivity ratio towards other PDE isoforms >1 were discarded
(see Table S2 in the Supplementary Material for values). Selectivity ratio was calculated
dividing estimated DG values for other PDE isoforms by DG values for PDE5 (2H44)
(Loza-Mej
ıa et al. 2018). The intersection between the resulting sets provided the final
group of molecules showing most promising calculated pharmacokinetic properties,
binding energy values and selectivity for PDE5 among the original array of 30 natural
compounds. This set of tentative leads includes apigenin, berberine, delphinidin, kraus-
sianone 2, 3 and 4, luteolin, naringenin, osajin and xanthohumol. The docking protocol
was preliminary assessed by comparing the computed results with the experimental
data on PDE5 inhibition and PDE-related vasorelaxant effects of the compounds avail-
able in the literature (see Table S3 in the Supplementary Material). Even though IC
50
values are not available for all the studied compounds, according to the retrieved data
the protocol was effective in predicting the potential efficacy of luteolin, osajin and
Figure 1. Chemical structures of some of the studied compounds belonging to the major investi-
gated chemical classes.
NATURAL PRODUCT RESEARCH 3

delphinidin and in correctly identifying most of non-active compounds (such
as quercetin and pomiferin). Although, the prediction of the behaviour of naringenin
and xanthohumol was less accurate. The reliability of the docking setup was also
verified by running the same in silico experiment using vardenafil, tadalafil, avanafil
and structures from the “Database of Useful Decoys: Enhanced”as ligands (Huang
et al. 2006; Mysinger et al. 2012), which showed estimated DG values ranging in
the higher half of the overall set (Table S4 in the Supplementary Material). Moreover,
from the point of view of the interaction motif, the accuracy of the prediction was
verified by re-docking the co-crystallized ligand sildenafil to PDE5 (2H42), obtaining
a superimposition between computed and experimental poses (Figure S1 in the
Supplementary Material).
Among the highlighted set of potentially bioactive natural molecules, kraussianone
2(Figure 1) showed the best estimated binding energy towards PDE5 (-11.9 kcal/mol).
This compound is contained in Eriosema kraussianum, which is used in South African
traditional medicine and is claimed to be effective against ED by Zulu traditional
health practitioners. Moreover, the vasorelaxant effect of kraussianones was previously
evaluated in rat models in comparison with sildenafil (Ojewole et al. 2006).
Kraussianone 2 has a low molecular weight (420.46 Da) and promising calculated
pharmacokinetic properties (TPSA ¼100.13 Å
2
and volume ¼376.57 Å
3
, in line with
sildenafil). In a more detailed docking study, the 3D binding motif of kraussianone 2
with PDE5 was compared to that computed for sildenafil. The compounds bind
to the same site and, as can be observed from 3D models and 2D interaction maps
Table 1. Estimated DG values (kcal/mol) for the studied compounds towards PDE isoforms. PDB
ID is given in parentheses.
Ligand
PDE5
(2H44)
PDE1
(5W6E)
PDE2
(4JIB)
PDE4
(1XOS)
PDE6
(3JWQ)
PDE7
(4PM0)
PDE9
(3JSW) PDE10 (5K9R)
Apigenin 10.6 9.7 9.4 9.6 8.6 9.6 8.8 8.3
Auriculasin 8.6 10.1 9.6 10.5 10.7 10.1 10.1 9.1
Berberine 11.7 10.1 9.7 10.0 9.5 7.3 9.4 9.0
Caffeic acid 7.4 6.9 6.8 7.1 7.2 6.9 6.6 6.3
Caffeine 7.3 5.9 6.3 6.0 6.1 6.3 5.8 6.5
Delphinidin 10.7 9.6 9.7 9.6 8.8 9.4 9.1 8.5
Forskolin 6.8 6.9 7.9 6.8 7.5 6.2 7.3 7.7
Gallic acid 7.2 5.9 5.8 6.1 5.9 6.2 5.9 5.8
Kraussianone 2 11.9 9.9 9.8 9.9 9.8 9.4 11.7 10.3
Kraussianone 3 11.8 8.8 9.8 9.9 9.8 9.6 11.7 10.3
Kraussianone 4 11.5 9.3 10.3 10.3 9.8 9.2 10.8 10.0
Kraussianone 5 8.5 8.8 9.4 8.7 8.7 8.6 9.0 8.1
Luteolin 10.6 9.7 9.6 10.0 9.2 10.0 9.0 8.4
Maca 2 5.4 6.3 6.7 6.4 5.9 6.7 6.1 6.0
Macaridine 7.4 6.9 6.9 7.0 7.5 7.3 7.2 6.2
Naringenin 10.3 9.6 9.2 9.6 8.6 9.6 8.6 8.2
Osajin 11.4 9.2 9.6 9.9 10.7 10.0 11.0 9.1
Osthol 9.4 7.3 7.6 8.0 7.7 7.1 8.0 7.4
Pomiferin 8.5 8.8 9.6 9.8 11.0 9.8 11.6 9.2
Quercetin 10.1 9.8 9.6 9.4 9.7 10.3 8.9 8.4
Resveratrol 9.3 8.4 8.5 8.8 8.0 8.5 8.2 7.4
Scandenone 8.1 10.1 9.3 10.7 10.3 10.6 10.0 8.9
Tyramine 6.1 5.9 5.6 6.0 6.1 6.1 5.3 5.1
Xanthohumol 10.2 8.8 8.9 8.9 8.5 9.2 8.3 8.5
Yohimbine 9.1 9.4 9.8 10.7 9.3 10.5 9.8 8.6
Sildenafil 10.1 8.2 9.3 10.3 9.8 9.4 9.7 7.8
4 A. ONGARO ET AL.

(Figures S2–S7 in the Supplementary Material), kraussianone 2 and the synthetic PDE5
inhibitor share common interaction patterns, mainly interfering with the same lipo-
philic residues (Val782, Phe786, Phe820) and some of the hydrophilic (His613, Asn661).
3. Conclusions
Therapeutic activity and adverse effects are consequences of sophisticated and inter-
connected molecular mechanisms related to the effective concentration reached by
the compounds in plasma and in tissues, where the several isoforms of PDEs are vari-
ously expressed. Drug-likeness and selectivity profile remain pivotal features for natural
molecules claimed to be effective on ED through PDE5 inhibition. Among a heteroge-
neous array of molecules, a group of natural compounds with different chemical scaf-
folds, satisfying predicted pharmacokinetic properties and good predicted selectivity
for PDE5 were highlighted. Kraussianone 2 proved to be the most promising com-
pound of the set according to the evaluated parameters.
Disclosure statement
The authors declare no conflict of interest.
Funding
This work was supported by University of Padova and University of Brescia.
References
Abusnina A, Lugnier C. 2017. Therapeutic potentials of natural compounds acting on cyclic
nucleotide phosphodiesterase families. Cell Signal. 39:55–65.
Arya H, Syed SB, Singh SS, Ampasala DR, Coumar MS. 2018. In silico investigations of chemical
constituents of Clerodendrum colebrookianum in the anti-hypertensive drug targets: ROCK,
ACE, and PDE5. Interdiscip Sci. 10(4):792–804.
Bischoff E. 2004. Potency, selectivity, and consequences of nonselectivity of PDE inhibition. Int J
Impot Res. 16 Suppl 1:S11–S14.
Chen CY. 2009. Computational screening and design of traditional Chinese medicine (TCM) to
block phosphodiesterase-5. J Mol Graph Model. 28(3):261–269.
Goldstein I, Lue TF, Padma-Nathan H, Rosen RC, Steers WD, Wicker PA. 1998. Oral sildenafil in
the treatment of erectile dysfunction. N Engl J Med. 338(20):1397–1404.
Goldstein I. 2000. Male sexual circuitry. working group for the study of central mechanisms in
erectile dysfunction. Sci Am. 283(2):70–75.
Huang N, Shoichet BK, Irwin JJ. 2006. Benchmarking sets for molecular docking. J Med Chem.
49(23):6789–6801.
Loza-Mej
ıa MA, Salazar JR, S
anchez-Tejeda JF. 2018. In silico studies on compounds derived
from calceolaria: phenylethanoid glycosides as potential multitarget inhibitors for the devel-
opment of pesticides. Biomolecules. 8(4):121.
Matsson P, Kihlberg J. 2017. How big is too big for cell permeability? J Med Chem. 60(5):
1662–1664.
Mysinger MM, Carchia M, Irwin JJ, Shoichet BK. 2012. Directory of useful decoys, enhanced
(DUD-E): better ligands and decoys for better benchmarking. J Med Chem. 55(14):6582–6594.
NATURAL PRODUCT RESEARCH 5

Petit J, Meurice N, Kaiser C, Maggiora G. 2012. Softening the Rule of Five-where to draw the
line? Bioorg Med Chem. 20(18):5343–5351.
Oboh G, Adebayo AA, Ademosun AO, Boligon AA. 2017. In vitro inhibition of phosphodiester-
ase-5 and arginase activities from rat penile tissue by two Nigerian herbs (Hunteria umbellata
and Anogeissus leiocarpus). J Basic Clin Physiol Pharmacol. 28(4):393–401.
Ojewole JA, Drewes SE, Khan F. 2006. Vasodilatory and hypoglycaemic effects of two pyrano-iso-
flavone extractives from Eriosema kraussianum N. E. Br. [Fabaceae] rootstock in experimental
rat models. Phytochemistry. 67(6):610–617.
Pavan V, Mucignat-Caretta C, Redaelli M, Ribaudo G, Zagotto G. 2015. The old made new: nat-
ural compounds against erectile dysfunction. Arch Pharm (Weinheim). 348(9):607–614.
Ribaudo G, Pagano MA, Pavan V, Redaelli M, Zorzan M, Pezzani R, Mucignat-Caretta C,
Vendrame T, Bova S, Zagotto G. 2015. Semi-synthetic derivatives of natural isoflavones from
Maclura pomifera as a novel class of PDE-5A inhibitors. Fitoterapia. 105:132–138.
Ribaudo G, Pagano MA, Bova S, Zagotto G. 2016. New therapeutic applications of phospho-
diesterase 5 inhibitors (PDE5-Is). Curr Med Chem. 23(12):1239–1249.
Ribaudo G, Vendrame T, Bova S. 2017. Isoflavones from Maclura pomifera: structural elucidation
and in silico evaluation of their interaction with PDE5. Nat Prod Res. 31(17):1988–1994.
Ribaudo G, Zanforlin E, Canton M, Bova S, Zagotto G. 2018. Preliminary studies of berberine and
its semi-synthetic derivatives as a promising class of multi-target anti-parkinson agents. Nat
Prod Res. 32(12):1395–1401.
Srivani P, Srinivas E, Raghu R, Sastry GN. 2007. Molecular modeling studies of pyridopurinone
derivatives-potential phosphodiesterase 5 inhibitors. J Mol Graph Model. 26(1):378–390.
T€
om€
ori T, Hajd
u I, Barna L, Lorincz Z, Cseh S, Dorm
an G. 2012. Combining 2D and 3D in silico
methods for rapid selection of potential PDE5 inhibitors from multimillion compounds’reposi-
tories: biological evaluation. Mol Divers. 16(1):59–72.
West E, Krychman M. 2015. Natural aphrodisiacs-A review of selected sexual enhancers. Sex Med
Rev. 3(4):279–288.
Yafi FA, Sharlip ID, Becher EF. 2018. Update on the safety of phosphodiesterase type 5 inhibitors
for the treatment of erectile dysfunction. Sex Med Rev. 6(2):242–252.
6 A. ONGARO ET AL.