Synthesis of new cyclic imides derived from safrole, structure-and ligand-based approaches to evaluate potential new multitarget agents against species of leishmania

Article abstract of DOI:10.2174/1573406415666190430144950

Background: Leishmaniasis is a neglected disease that does not have adequate treatment. It affects around 12 million people around the world and is classified as a neglected disease by the World Health Organization. In this context, strategies to obtain new, more active and less toxic drugs should be stimulated. Sources of natural products combined with synthetic and chemoinformatic methodologies are strategies used to obtain molecules that are most likely to be effective against a specific disease. Computer-Aided Drug Design has become an indispensable tool in the pharmaceutical industry and academia in recent years and has been employed during various stages of the drug design process. Objectives: Perform structure-and ligand-based approaches, synthesize and characterize some compounds with materials available in our laboratories to verify the method’s efficiency. Methods: We created a database with 33 cyclic imides and evaluated their potential anti-Leishmanial activity (L. amazonensis and L. donovani) through ligand-and structure-based virtual screening. A diverse set selected from ChEMBL databanks of 818 structures (L. donovani) and 722 structures (L. amazonensis), with tested anti-Leishmanial activity against promastigotes forms, were classified according to pIC₅₀ values to generate and validate a Random Forest model that shows higher statistical indices values. The structures of four different L. donovani enzymes were downloaded from the Protein Data Bank and the imides’ structures were submitted to molecular docking. So, with available materials and technical feasibility of our laboratories, we have synthesized and characterized seven compounds through cyclization reactions between isosafrole and maleic anhydride followed by treatment with different amines to obtain new cyclic imides to evaluate their anti-Leishmanial activity. Results: In silico study allowed us to suggest that the cyclic imides 5_{16, 25, 31, 24, 32, 2, 3, 22} can be tested as potential multitarget molecules for leishmanial treatment, presenting activity probability against four strategic enzymes (Topoisomerase I, N-myristoyltransferase, cyclophilin and O-acetylserine sulfhydrylase). The compounds synthesized and tested presented pIC₅₀ values less than 4.7 for Leishmania amazonensis. Conclusion: After combined approach evaluation, we have synthesized and characterized seven cyclic imides by IR,¹ H NMR,¹³ C-APT NMR, COSY, HETCOR and HMBC. The compounds tested against promastigote forms of L. amazonensis presented pIC₅₀ values less than 4.7, showing that our method was efficient in predicting true negative molecules.

Full text of DOI:10.2174/1573406415666190430144950

39
Send Orders for Reprints to reprints@benthamscience.net
Medicinal Chemistry, 2020, 16, 39-51
RESEARCH ARTICLE
ISSN: 1573-4064
eISSN: 1875-6638
Impact
Factor:
2.53
Synthesis of New Cyclic Imides Derived from Safrole, Structure- and
Ligand-based Approaches to Evaluate Potential New Multitarget Agents
Against Species of Leishmania
José A. de Sousa Luis1,4,*, Normando A. da Silva Costa2, Cristiane C.S. Luis3, Bruno F. Lira2,
Petrônio F. Athayde-Filho2, Tatjana K. de Souza Lima4, Juliana da Câmara Rocha4, Luciana Scotti4
and Marcus T. Scotti4,*
2
1Education and Health Center, Federal University of Campina Grande, Cuité, Paraíba, Brazil; Research Laboratory
on Biofuels and Organic Synthesis, Chemistry Department, Federal University of Paraíba, João Pessoa, Paraíba, Bra-
3
zil; Postgraduate Program in Nutrition Sciences, Federal University of Paraíba, João Pessoa, Paraíba, Brazil;
4Postgraduate Program in Natural Products and Synthetic Bioactive, Federal University of Paraíba, João Pessoa,
Paraíba, Brazil
Abstract: Background: Leishmaniasis is a neglected disease that does not have adequate treat-
ment. It affects around 12 million people around the world and is classified as a neglected disease
by the World Health Organization. In this context, strategies to obtain new, more active and less
toxic drugs should be stimulated. Sources of natural products combined with synthetic and
chemoinformatic methodologies are strategies used to obtain molecules that are most likely to be
effective against a specific disease. Computer-Aided Drug Design has become an indispensable
tool in the pharmaceutical industry and academia in recent years and has been employed during
various stages of the drug design process.
Objectives: Perform structure- and ligand-based approaches, synthesize and characterize some
compounds with materials available in our laboratories to verify the method’s efficiency.
A R T I C L E H I S T O R Y
Methods: We created a database with 33 cyclic imides and evaluated their potential anti-
Leishmanial activity (L. amazonensis and L. donovani) through ligand- and structure-based virtual
screening. A diverse set selected from ChEMBL databanks of 818 structures (L. donovani) and
722 structures (L. amazonensis), with tested anti-Leishmanial activity against promastigotes forms,
were classified according to pIC50 values to generate and validate a Random Forest model that
shows higher statistical indices values. The structures of four different L. donovani enzymes were
downloaded from the Protein Data Bank and the imides’ structures were submitted to molecular
docking. So, with available materials and technical feasibility of our laboratories, we have synthe-
sized and characterized seven compounds through cyclization reactions between isosafrole and
maleic anhydride followed by treatment with different amines to obtain new cyclic imides to
evaluate their anti-Leishmanial activity.
Received: April 28, 2018
Revised: February 12, 2019
Accepted: April 15, 2019
DOI:
10.2174/1573406415666190430144950
Results: In silico study allowed us to suggest that the cyclic imides 516, 25, 31, 24, 32, 2, 3, 22 can be
tested as potential multitarget molecules for leishmanial treatment, presenting activity probability
against four strategic enzymes (Topoisomerase I, N-myristoyltransferase, cyclophilin and O-
acetylserine sulfhydrylase). The compounds synthesized and tested presented pIC50 values less
than 4.7 for Leishmania amazonensis.
Conclusion: After combined approach evaluation, we have synthesized and characterized seven
1
cyclic imides by IR, H NMR, 13C-APT NMR, COSY, HETCOR and HMBC. The compounds
tested against promastigote forms of L. amazonensis presented pIC50 values less than 4.7, showing
that our method was efficient in predicting true negative molecules.
Keywords: Cyclic imides, safrole, virtual screening, molecular docking, Leishmania amazonensis, Leishmania donovani, anti-
leishmania activity.
1. INTRODUCTION
*Address correspondence to these authors at the Postgraduate Program
in Natural Products and Synthetic Bioactive, Federal University of
Despite the great development of modern medicinal
Paraíba, João Pessoa, Paraíba, Brazil; Tel: +55 83 99869-0415;
chemistry, there are some microbial diseases that remain
E-mails: mtscotti@gmail.com, jalixluis@hotmail.com
1875-6638/20 $65.00+.00
© 2020 Bentham Science Publishers

40 Medicinal Chemistry, 2020, Vol. 16, No. 1
Luis et al.
without adequate chemotherapeutic agents, either due to
problems of toxicity or resistance, among them is Leishma-
niasis, which is a complex of infectious diseases caused by
parasites of the family Trypanossomatidae and genus
Leishmania [1, 2]. It affects around 12 million people around
the world; there are reported cases in 98 countries spread
across five continents, mainly in poor countries, causing the
disease to be classified as a neglected disease by the World
Health Organization [2-4]. It is the second most parasitic
disease with the highest mortality and an estimated 350 mil-
lion people are at risk of infection, being more prevalent in
Brazil, Sudan, Ethiopia and India [5]. The protozoan
Leishmania is transmitted to humans by insects of the genus
Phlebotomus spp. and Lutzomya spp. and diseases caused by
infection are classified as Visceral Leishmaniasis and Cuta-
neous Leishmaniasis [6]. Another problem that has arisen
recently is Leishmania coinfection in patients with Human
Immunodeficiency Virus, mainly due to the development of
resistance in the parasites to drugs commonly used in ther-
apy (the pentavalent antimonial drugs Pentostan and Glucan-
time) [7].
Taking into account the great medicinal importance of
cyclic imides and their derivatives, as well as the excellent
results found and published to date, the planning of 33 cyclic
imides was carried out and a combined ligand- and structure-
virtual chemical screening was carried out to select mole-
cules with higher probability to show the desired effect
against selected Leishmania targets. The selected compounds
were synthesized and their anti-Leishmanial activities were
evaluated.
2. MATERIAL AND METHODS
2.1. Dataset
We selected a diverse set of 818 structures from the
ChEMBL database, (https://www.ebi.ac.uk/chembl/), which
had been screened (in vitro) to inhibit the promastigote L.
donovani and another diverse set of 722 structures which had
been screened for promastigote form of L. amazonensis. The
compounds were classified using values of -logIC50 (mol/L)
= pIC50, which led us to assign 293 actives (pIC50 ≥ 5.0) and
525 inactives (pIC50 ≤ 4.7) for L. donovani; 253 actives
(pIC50 ≥ 5.0) and 469 inactives (pIC50 ≤ 4.7) for L. ama-
zonensis. We used a border in the pIC50 values looking for
better prediction results. In this case, IC50 represented the
concentration required for 50% inhibition of promastigote L.
donovani and L. amazonensis. The compounds with pIC50
values between 4.7 and 5.0 were excluded to minimize
the border effect and improve the discriminant power
of the generated models. Our databank includes compounds
In this context, strategies to obtain new, more active and
less toxic drugs should be stimulated. Sources of natural
products combined with synthetic and chemoinformatic
methodologies are strategies used to obtain molecules that
are most likely to be effective against a specific disease.
Computer-Aided Drug Design has become an indispensable
tool in the pharmaceutical industry and academia in recent
years and has been employed during various stages of the
drug design process. Initially, this method focuses on reduc-
ing the overall number of possible ligands; in the later
stages, during lead optimization, the emphasis shifts to re-
ducing experimental costs and the duration of time required
to make a discovery [8-10].
51-33
. For all structures, SMILES codes were used
as input data to Marvin 17.18.0.1784, 2017, ChemAxon
(http://www.chemaxon.com). We used Standardizer
software [JChem 17.18.0.1784, 2017; ChemAxon
(http://www.chemaxon.com)] to canonize structures, add
hydrogens, perform aromatic form conversions, clean the
molecular graph in three dimensions and save compounds in
sdf format [9, 19].
Cyclic imides are a group of compounds with various
biological activities that present a large class of compounds
obtained by organic synthesis including several subclasses,
among them maleimides, succinimides, glutarimides,
phthalimides and naphthalimides, as well as their respective
derivatives [11]. Because they are electronically neutral and
of a hydrophobic nature, they easily cross cell membranes,
leading to the important pharmacological effects of these
imides, such as anti-inflammatory, antitumor, antimicrobial
activities among others, which may be related to the size and
characteristics of the groups present in the imidic ring, which
may alter the steric characteristics of the molecules, altering
their activity [12-16]. Another important class of substances
for the pharmaceutical industry is the essential oils, which
are natural products, which present as aromatic and oily liq-
uids, evaporating easily when exposed to air at room tem-
perature. Because of this, they are also known as volatile
oils, other names are ethereal, refringent and essential oils.
These oils are formed in several plants as byproducts of sec-
ondary metabolism [17]. These oils do not present as pure
mixtures but rather mixtures with various proportions of dif-
ferent chemical structures such as acids, aldehydes, alcohols,
ketones, esters, ethers, phenols, aromatic or terpene hydro-
carbons [18]. As an example, we have sassafras oil (Ocotea
pretiosa), which is rich in safrole, the precursor used in this
study, because it has the methylenedioxy group that provides
important characteristics to the derived molecules.
2.2. Volsurf Descriptors
3-D structures were used as input data in the Volsurf+
program v. 1.0.7 and were subjected to molecular interaction
fields (MIFs) to generate descriptors using the following
probes: N1 (amide nitrogen-hydrogen bond donor probe), O
(carbonyl oxygen-hydrogen bond acceptor probe), OH2 (wa-
ter probe) and DRY (hydrophobic probe) [20]. Additional
non-MIF-derived descriptors were generated to create a total
of 128 descriptors. Volsurf descriptors have been previously
used to predict anti-Leishmanial activity of natural products
on enzymes and predict the activity of some molecules [21,
22].
2.3. Models
KNIME 3.4.0 software (KNIME 3.4.0 the Konstanz In-
formation Miner Copyright, 2003-2017), (www.knime.org)
[23] was used to perform all of the following analyses. The
descriptors and class variables were imported from the Vol-
surf+ program, v. 1.0.7, and the data were divided using the
“Partitioning” node with the “stratified sample” option to
create a training set and a test set, encompassing 80% and
20% of the compounds, respectively. Although the com-
pounds were selected randomly, the same proportion of ac-

Synthesis of New Cyclic Imides Derived from Safrole, Structure- and Ligand-based
Medicinal Chemistry, 2020, Vol. 16, No. 1 41
1
tive and inactive samples was maintained in both sets. For
internal validation, we employed cross-validation using 10
randomly selected, stratified groups, and the distributions
according to activity class variables were found to be main-
tained in all validation groups and in the training set. De-
scriptors were selected, and a model was generated using the
training set and the RF algorithm [24], using the WEKA
nodes [25]. The parameters selected for RF included the fol-
lowing settings: number of trees to build = 1900, seed for
random number generator = 1909501934341. The internal
and external performances of the selected models were ana-
lyzed for sensitivity (true positive rate, i.e., active rate),
specificity (true negative rate, i.e., inactive rate) and accu-
racy (overall predictability). In addition, the sensitivity and
specificity of the ROC curve were found to describe the true
performance with more clarity and accuracy. The plotted
ROC curve shows the true positive (active) rate either versus
the false positive rates or versus sensitivity (1: specificity).
In a two-class classification, when a variable that is being
investigated cannot be distinguished between the two groups
(i.e., when there is no difference between the two distribu-
tions), the area under the ROC curve equals 0.5, which is to
say that the ROC curve will coincide with the diagonal.
When there is a perfect separation of values between the two
groups (i.e., no overlapping of distributions), the area under
the ROC curve equals 1, which is to say that the ROC curve
will reach the upper left corner of the plot [26].
MERCURY apparatus at 60, 200 and 500 MHz for H and
15, 50 and 125 MHz for 13C (Analytical Center-CBIOTEC-
UFPB), using tetramethylsilane (TMS) as the internal refer-
ence and the solvents dimethyl sulfoxide (DMSO-d6) and
chloroform (CDCl3) to solubilize the samples. The chemical
shifts (δ) were measured in units of parts per million (ppm)
and the coupling constants in hertz (Hz). The melting points
were determined on a heating plate MQAPF-3 (LPBS-
DQ/UFPB) and were not corrected.
2.5.1. Preparation of Isosafrole (6,7-methylenedioxypro-
penylbenzene) (2)
10 g (62 mmol, 9.1 mL) of safrole (1) and 50 mL (150
mmol) of a 3 M solution of KOH in n-butanol were added to
a 125 mL flask equipped with reflux condenser and magnetic
stirrer. The reaction was kept under stirring at reflux for 6 h.
After this time, the mixture was neutralized with 10% HCl
and the organic phase was washed successively with distilled
water and aqueous NaCl solution to provide 9.5 g (8.65 mL,
1
95%) of isosafrole (2) as a colorless liquid. H NMR, (60
MHz, DMSO-d6, δ): 1.82 (d, 3H), 5.82 (s, 2H), 6.73-6.92 (m,
5H). 13C NMR, (15 MHz, DMSO-d6, δ): 148.2, 146.8, 132.2,
130.5, 123.3, 120.2, 108.0, 105.4, 100.9, 18.1.
2.5.2. Preparation of 11,12-methylenedioxy-5-methyl-
3,4,5,6-tetrahydronaphthalene-2,15-dicarboxylic
dride (4)
Anhy-
A mixture containing isosafrole (2) (13 g, 80 mmol),
maleic anhydride (3) (10 g, 101 mmol - excess) and xylene
(40 mL) was refluxed for about 3 h at a temperature of about
100 °C. After reflux the reaction mixture was cooled and
precipitated, the solid obtained was washed with ethanol and
extracted with hot chloroform to afford 9.36 g of (4) as pale-
yellow crystals. Yield: 46%. Mp: 141°C -literature 142-
143°C [32]. IR (KBr, ν cm-1): 2978, 2935, 2902 (C-H) 1788,
1726 (C=O); 1502, 1483 (C=C); 1388 (CH); 1236, 1033 (C-
O); 910, 866, 756 (aromatic). 1H NMR, (60 MHz, DMSO-d6,
δ): 1.06 (s, 1H); 2.59 (m, 3H); 3.68 (dd, 3H); 4.46 (d, 1H);
5.96 (t, 2H); 6.68 (s, 1H); 6.97 (s, 1H).
2.4. Docking
The structure of L. donovani enzymes TOPI [27], NMT
[28], Cyp [29] and OASS [30] downloaded from the
Protein Data Bank (PDB) are summarized in Table 5
(http://www.rcsb.org/pdb/home/home.do). 51-33 structures
were submitted to molecular docking using the Molegro Vir-
tual Docker, v. 6.0.1 (MVD). All of the water molecules
were deleted from the enzyme structure, and the enzyme and
compound structures were prepared using the same default
parameter settings in the same software package. The dock-
ing procedure was performed using a GRID of 15 Å in ra-
dius and 0.30 Å in resolution to cover the ligand-binding site
of the enzyme’s structures. The Moldock score algorithm
was used as the score function [31]. For all enzymes, the
binding site was the same as the ligand present in the PDB
file.
2.5.3. General Procedure for the Preparation of the Cyclic
Imides
The substituted aniline (1.5 mmol) and (4) (1.5 mmol) in
3.0 mL acetic acid (or amount needed to solubilize the solids
well) were dissolved in a flask, the reaction mixture was
refluxed for about 3 h after cooling the solution, the reaction
mixture was allowed to stand until precipitation, filtered and
washed with distilled water and the solid obtained was re-
crystallized from ethanol.
2.5. Chemistry
Reagents used for the synthesis: maleic anhydride
(Merck, 99%), potassium hydroxide (Vetec, 85%), pheny-
lamine (Aldrich, 99.5%), benzylamine (Riedel, 99%), 4-
aminobenzoic acid (Vetec, 99%), sulfanilamide (Vetec,
99%), 4-chloro-3-nitroaniline (Richem, 98%), 4-bromo-3-
nitroaniline (Richem, 99%), 4-fluoro-3-nitroaniline (Richem,
99%) and safrole (Aldrich, 97%). Solvents used for the syn-
thesis of the compounds: water, chloroform (Merck), deuter-
ated chloroform (Isotec), DMSO (Vetec), ethanol (Merck),
methanol (Merck), acetic acid (Merck), n-butanol, dichloro-
methane and toluene (Vetec). Infrared spectra were recorded
on a BOMEM spectrometer MB100 model M Series (LA-
SOM - CCEN - UFPB) on KBr pellets. Absorption bands are
expressed in cm-1, in the range of 4000-400 cm-1. 1-D and 2-
2.5.3.1. Preparation of 1-(19-fluoro-18-nitrophenyl)-11,12-
methylenedioxy-5-methyl-3,4,5,6-tetrahydronaphthalene-
2,15-dicarboxyimide (51)
Compound 51 was obtained as pale-yellow crystals.
Yield: 70%. Mp: 180-182 °C. IR (KBr, ν, cm-1) 3082 (C-H,
aromatic); 2962, 2920, 2877 (C-H, aliphatic); 1707 (C=O)
1
1502, 1483 (C=C), aromatic). H NMR (200 MHz, DMSO-
d6, δ) 3,52 (dd, J = 8.9, 5.7 Hz, 1H); 4,27 (d, J = 9.0 Hz, 1H);
2,26 (m, 1H), 2,61 (m, 2H); 6,74 (s, 1H); 7,09 (s, 1H); 5,98
(d, J = 15,8 Hz, 2H); 1,13 (d, J = 7.2 Hz, 3H); 8,17 (d, J =
2.0 Hz, 1H); 7,78-7,70 (m, 1H); 7,78 - 7,70 (m, 1H). 13C
1
D H and 13C NMR spectra were obtained on a VARIAN

42 Medicinal Chemistry, 2020, Vol. 16, No. 1
Luis et al.
NMR (50 MHz, DMSO-d6, δ) 176.5, 175.9, 153.9, 146.3,
145.7, 136.7, 135.0, 130.2, 128.6, 124.65, 122.1, 119.1,
109.3, 108.6, 100.8, 43.9, 43.4, 34.5, 29.9, 16.5.
(d, J = 8.6 Hz, 1H); 7,47 (d, J = 8.6 Hz,1H); 7,47 (d, J = 8.6
Hz, 1H); 8,93 (d, J = 8.6 Hz, 1H); 7,48 (s, 2H - NH2). 13C
NMR (50 MHz, DMSO-d6, δ) 176.9, 176.2, 146.4, 145.8,
143.9, 135.1, 129.9, 127.7, 127.7, 126.6, 126.6, 122.3, 109.3,
108.8, 100.9, 43.9, 43.3, 34.7, 29.8, 16.4.
2.5.3.2. Preparation of 1-(19-bromo-18-nitrophenyl)-11,12-
methylenedioxy-5-methyl-3,4,5,6-tetrahydronaphthalene-
2,15-dicarboxyimide (52)
2.5.3.6. Preparation of 1-benzyl-11,12-methylenedioxy-5-
methyl-3,4,5,6-tetrahydronaphthalene-2,15-dicarboxyimide
(56)
Compound 52 was obtained as yellow crystals. Yield:
68%. Mp: 192-195 °C. IR (KBr, ν, cm-1) 3086 (C-H, aro-
matic); 2956, 2927, 2906, 2873 (C-H, aliphatic); 2852, 2821
Compound 56 was obtained as pale-yellow crystals. Yi-
eld: 82.7%. Mp: 139-140 °C. IR (KBr, ν, cm-1) 2966, 2912
(C-H, aliphatic); 1693 (C-O); 1504, 1485 (C=C, aromatic);
1398 (CH3). 1H NMR (200 MHz, DMSO-d6, δ) 3,36 (dd, J =
8.9, 5.8 Hz, 1H); 4,14 (d, J = 8.7 Hz, 1H); 2,23 (m, 1H); 2,52
(m, 2H); 6,70 (s, 1H) 7,07 (s, 1H); 5,97 (d, J = 7.3 Hz, 2H);
0,96 (d, J = 6.3 Hz, 3H); 7,23 (m, 5H); 4,55 (s, 2H). 13C
NMR (50 MHz, DMSO-d6, δ) 177.7, 177.1, 146.2, 145.7,
136.2, 129.9, 128.6, 128.6, 128.4, 127.5, 127.4, 122.6, 109.2,
108.6, 100.9, 46.6, 43.2, 41.5, 34.6, 29.9, 16.5.
1
(C-H, aliphatic); 1710 (C=O); 1541 (C=C, aromatic). H
NMR (200 MHz, DMSO-d6, δ) 3,50 (dd, J = 9.1, 5.5 Hz,
1H); 4,25 (d, 1H); 2,22 (m, 1H); 2,45 (m, 2H); 6,72 (s, 1H);
7,06 (s, 1H); 5,96 (d, J = 18.8 Hz, 2H), 1,11 (d, J = 7.0 Hz,
3H); 8,05 (d, J = 7.3Hz, 1H); 8,03 (d, J = 7.3, 1H); 7,55 (dd,
J = 8.6, 2.4 Hz, 1H). 13C NMR (50 MHz, DMSO-d6, δ)
176.4, 175.8, 149.6, 146.4, 145.8, 132.4, 132.2, 130.3, 124.0,
123.9, 122.1, 112.9, 109.3, 108.7, 100.9, 43.9, 43.5, 34.6,
29.9, 16.7.
2.5.3.3. Preparation of 1-(19-chloro-18-nitrophenyl)-11,12-
methylenedioxy-5-methyl-3,4,5,6-tetrahydronaphthalene-
2,15-dicarboxyimide (53)
2.5.3.7. Preparation of 1-phenyl-11,12-methylenedioxy-5-
methyl-3,4,5,6-tetrahydronaphthalene-2,15-dicarboxyimide
(57)
Compound 53 was obtained as pale-yellow crystals. Yi-
eld: 72.8%. Mp: 198-200 °C. IR (KBr, ν, cm-1) 2958, 2929,
2910 (C-H, aliphatic); 1708 (C=O); 1541, 1479 (C=C, anel
Compound 57 was obtained as pale-yellow crystals.
Yield: 82%. Mp: 246-248 °C-literature 249 °C. IR (KBr, ν,
cm-1) 3093 (C-H, aromatic); 2922, 2982, 2841 (C-H, alipha-
tic); 1681 (C=O); 1591, 1573 (C=C, aromatic); 1424 (CH3);
1
aromatic); 1388 (CH3). H NMR (200 MHz, DMSO-d6, δ)
1
3,52 (dd, J = 8.9, 5.7 Hz, 1H); 4,28 (d. J = 8.9 Hz, 1H); 2,25
(m, 1H); 2,60 (m, 2H) 6,75 (s, 1H); 7,08 (s, 1H); 5,98 (d, J =
17.8 Hz, 2H); 1,08 (d, J = 7.0 Hz, 3H); 8,12 (d, J = 2.4 Hz,
1H); 7, 92 (d, J = 8.7 Hz,1H); 7, 68 (dd, J = 8.7, 2.4 Hz, 1H).
13C NMR (50 MHz, DMSO-d6, δ) 176.4, 175.8, 147.5,
146.4, 145.8, 132.4, 131.9, 130.3, 124.8, 124.2, 124.0, 122.1,
109.3, 108.7, 100.9, 43.9, 43.5, 34.6, 29.9, 16.7.
1321, 1091 (C-O). H NMR (200 MHz, DMSO-d6, δ) 3,49
(d, J = 8.9 Hz, 1H); 4,25 (d, J = 8.9 Hz0 1H); 2,29 (m, 1H);
2,61 (m, 2H); 5.99 (d, J = 9.7 Hz, 2H); 6,75 (s, 1H); 7,10 (s,
1H); 7,48 (t, J = 7.8 Hz, 2H); 7,22 (dd, J = 8.9, 1,5 Hz, 2H);
7,41 (t, J = 7.9 Hz, 1H); 7,22 (dd, J = 8.9, 1,5 Hz, 2H). 13C
NMR (50 MHz, DMSO-d6, δ) 177.0, 176.4, 146.3, 145.7,
132.3, 129.7, 128.9, 128.9, 128.4, 127.0, 127.0, 122.4, 109.3,
108.7, 100.8, 43.8, 43.0, 34.7, 29.7, 16.3.
2.5.3.4. Preparation of para-(11,12-methylenedioxy-
5-methyl-3,4,5,6-tetrahydronaphthalene-2,15-icarboxyimi-
de)-benzoic Acid (54)
2.6. Biological Activity
The parasite Leishmania (Leishmania) amazonensis
(IFLA/BR/1967/PH8) in promastigote form was cultivated in
Schneider’s medium (Sigma-Aldrich, St. Louis, MO, USA)
supplemented with 20% bovine fetal serum, 1% male human
urine and antibiotics (penicillin 200 U/mL and streptomycin
0.1 mg/mL) (Gibco, CA) kept incubated in a biological
oxygen demand (B.O.D.) incubator at 26°C. The colorimet-
ric assay agent MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-
diphenyltetrazolium bromide) is based on the dehydrogenase
activity of cytosolic and mitochondrial enzymes that are able
to reduce MTT (yellow coloration) and form a blue product,
insoluble in water, a formazan salt. The MTT assay was used
to evaluate the anti-Leishmania activity of compounds 51-7,
with determination of cell viability by calculating the IC50. In
a 96-well plate was added 100 µL supplemented Schneider
medium and about 1 × 106 promastigotes/well of L. ama-
zonensis. Subsequently, they were added to triplicate test
substances previously diluted in supplemented Schneider
medium to a final volume of 100 µL in each well at concen-
trations of 100, 50, 25, 12.5, 6.25 and 3.125 µg/mL of each
compound previously diluted in half with Schneider DMSO.
Then, the plate was incubated for 72 h in a B.O.D. oven at
26°C. At the end of incubation, 10 µL of MTT diluted in
PBS was added to a final concentration of 5 mg/mL. They
Compound 54 was obtained as brown crystals. Yield:
72.5%. Mp: 214 °C. IR (KBr, ν, cm-1) 3458 (OH); 2964,
1
2924, 2902 (C-H, aliphatic); 1708 (C=O); 1382 (CH3). H
NMR (200 MHz, DMSO-d6, δ) 3,56 (dd, J = 8.8, 5.6 Hz,
1H); 4,27 (d, J = 8.8 Hz, 1H); 2,27 (m, 1H); 2,57 (m, 2H);
6,74 (s, 1H); 7,08 (s, 1H); 5,98 (d, J = 15.6 Hz, 2H); 1,06 (d,
J = 7.0 Hz, 3H); 7,39 (d, J = 8.4 Hz, 1H); 8,03 (d, J = 8.4
Hz, 1H); 8,03 (d, J = 8.4 Hz, 1H); 7,39 (d, J = 8.4 Hz, 1H).
13C NMR (50 MHz, DMSO-d6, δ) 176.2, 176.1, 166.6,
146.3, 145.7, 136.0, 130.1, 129.9, 129.9, 129.8, 127.0, 127.0,
122.2, 109.3, 108.7, 100.7, 43.9, 43.2, 34.6, 29.7, 16.3.
2.5.3.5. Preparation of 1-(11,12-methylenedioxy-5-methyl-
3,4,5,6-tetrahydronaphthalene-2,15-dicarboxyimide)-p-
benzenesulfonylamine (55)
Compound 55 was obtained as pale-yellow crystals. Yi-
eld: 75%. Mp: 280 °C. IR (KBr, ν, cm-1) 3387, 3240 (N-H);
3105 (C-H, aromatic); 2974, 2916 (C-H, aliphatic); 1695
(C=O); 1595 (NH2); 1500, 1485 (C=C, aromatic), 1394
1
(CH), 1336, 1035 (S=O); 1035 (S=O). H NMR (200 MHz,
DMSO-d6, δ) 3.52 (dd, J = 8.8, 5.8 Hz, 1H), 4,29 (d, J = 8.9
Hz, 1H); 2,28 (m, 1H); 2,61 (m, 2H); 6,77 (s, 1H); 7,10 (s,
1H); 5,98 (d, J = 7.7 Hz, 2H); 1,09 (d, J = 6.8 Hz 3H); 8, 93

Synthesis of New Cyclic Imides Derived from Safrole, Structure- and Ligand-based
Medicinal Chemistry, 2020, Vol. 16, No. 1 43
Table 1. Summary of training, internal cross-validation, test results and corresponding match results, which were obtained using
the RF algorithm on the total set of 722 compounds tested against promastigotes forms of L. amazonensis (614 were in the
training set and 108 in the test set).
-
Validation
Match
Test
Samples
% Match
Samples
Match
% Match
Active
Inactive
Overall
253
469
722
215
399
614
70.23
90.22
83.22
38
70
25
67
92
65.79
95.71
85.19
108
Table 2. Summary of training, internal cross-validation, test results and corresponding match results, which were obtained using
the RF algorithm on the total set of 818 compounds tested against promastigotes forms of L. donovani (655 were in the
training set and 163 in the test set).
-
Validation
Match
Test
Samples
% Match
Samples
Match
% Match
Active
Inactive
Overall
293
525
818
235
420
655
71.06
91.19
83.97
58
46
96
79.31
91.43
87.12
105
163
142
were incubated for another 4 h in a B.O.D. greenhouse at
26°C, and then 50 µL 10% sodium dodecylsulfate (Vetec®)
was added. The plate was left overnight for dissolution of the
formazan, and finally the absorbance of each well was read
using a spectrophotometer (Spectramax Plus, Molecular De-
vices, Sunnyvale, CA, USA) at 570 nm. The negative control
was performed in Schneider medium supplemented with
0.2% DMSO. The positive control was performed in the
presence of amphotericin B as the reference drug. Assays
were performed in triplicate and repeated three times on dif-
ferent days.
were better at predicting the inactive compounds in both
studies (Tables 1 and 2); therefore, the specificity (true posi-
tive rate) was lower for the cross-validation and test sets
(70.23% and 65.79%, respectively) than the sensitivity (true
negative rate), which were 90.22% and 95.71%, respectively,
in the model for L. amazonensis (Table 1). In the model for
L. donovani, the specificities for the cross-validation and test
sets were 71.06% and 79.31%, respectively, and the sensi-
tivities were 91.19% and 91.43%, respectively (Table 2).
For both RF models, L. amazonensis and L. donovani, the
ROC plot that was generated for the test set, which plotted
the true positive (active) rate against the false positive rates
had an area under the curve (AUC) value of approximately
0.88 and 0.91, respectively, which is significantly higher
than 0.5 (Fig. 1). The Matthews Correlation Coefficient
(MCC) values for cross-validation and test sets were 0.624
and 0.670, for the L. amazonensis RF model and 0.645 and
0.716 for the L. donovani RF model. Because an MCC value
of 1 represents a perfect prediction, 0 represents random pre-
diction, and -1 represents total disagreement between predic-
tion and observation, the RF model shows significant MCC
values.
3. RESULTS AND DISCUSSIONS
3.1. Computational Study
3.1.1. Ligand-based VS Approach
The Volsurf (v 1.0.7) program generated 128 descriptors
that, together with the dependent variables (binary classifica-
tion) that described whether the compounds were active (A)
or inactive (I), were used as input data in the KNIME pro-
gram (v. 3.4.0) to generate the Random Forest (RF) model.
For all compounds, the generation of all 128 descriptors by
Volsurf+ was rapid, taking approximately 25 minutes using a
computer with an i7 processor, running at 2.6 GHz, and
equipped with 8 GB of RAM.
Table 3 shows a prediction of activity, with their respec-
tive probabilities (p) calculated using the RF model for com-
pounds 51-33 against L. amazonensis and L. donovani. For
values of p ≥ 0.5 the molecule was considered active (A); p
< 0.5 the molecule was considered inactive (I).
Table 1 summarizes the statistical indices of the RF
model for the training, cross-validation and test sets for
compounds tested against promastigotes forms of L. ama-
zonensis and Table 2 for compounds tested against promas-
tigotes forms of L. donovani. For the training set, the learn-
ing machine model gave the same hit rates for the inactive
compounds and active compounds, which were 100%. How-
ever, for the cross-validation and test sets, the match rates
On the other hand, we evaluated the potential of synthe-
sized imides as anti-Leishmanial leads against L. donovani
using the RF model and docking on selected Leishmania
donovani enzyme targets because only data deposited for this
species were found in the Protein Data Bank (Tables 4, 5).

44 Medicinal Chemistry, 2020, Vol. 16, No. 1
Luis et al.
(A)
(B)
1.000
1.000
0.900
0.800
0.700
0.600
0.900
0.800
0.700
0.600
0.500
0.400
0.300
0.200
0.500
0.400
0.300
0.200
0.100
0
0.100
0
AUC=0.91
0.900
1.000
AUC=0.88
0.100
0.300
0.500
0.700
0.100
0.300
0.500
0.700
0.900
0.800 1.000
0
0.200
0.400
0.600
0.800
0
0.200
0.400
0.600
Fig. (1). ROCs plots generated by the selected RF models for the test set and values of the area under the curve (AUC). (a) For L.
amazonensis and (b) for L. donovani. (A higher resolution / colour version of this figure is available in the electronic copy of the article).
Table 3. Prediction of activity, with their respective probabilities (p) calculated using the RF model for compounds 51-33 against L.
donovani and L. amazonensis
L. donovani
Activity
L. amazonensis
Activity
Compound
p
p
51
52
A
A
A
I
0.78
0.81
0.81
0.37
0.37
0.51
0.75
0.62
0.58
0.51
0.52
0.59
0.54
0.66
0.69
0.82
0.39
0.61
0.47
0.51
0.46
0.68
0.48
0.71
0.75
0.39
0.66
0.64
0.65
0.63
0.71
0.53
0.53
I
I
0.39
0.47
0.43
0.26
0.32
0.37
0.35
0.28
0.32
0.59
0.30
0.30
0.40
0.29
0.35
0.43
0.29
0.39
0.37
0.31
0.30
0.51
0.31
0.56
0.61
0.25
0.39
0.45
0.36
0.63
0.62
0.62
0.64
53
I
54
I
55
I
I
56
A
A
A
A
A
A
A
A
A
A
A
I
I
57
I
58
I
59
I
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
A
I
I
I
I
I
I
I
A
I
I
I
A
I
I
I
A
I
A
I
A
A
I
A
A
I
A
A
A
A
A
A
A
I
I
I
A
A
A
A

Synthesis of New Cyclic Imides Derived from Safrole, Structure- and Ligand-based
Medicinal Chemistry, 2020, Vol. 16, No. 1 45
Table 4. Data of the tested enzymes.
Enzyme
PDB ID
Resolution (Å)
Source
Ligand
TOPI
NMT
Cyp
2B9S
5A27
3EOV
3T4P
2.27
1.42
2.6
L. donovani
L. donovani
L. donovani
L. donovani
VO4
2-oxopentadecyl-CoA
cyclosporin A
OASS
1.79
Di(hydroxyethyl) ether
Table 5. Moldock score of best position of imides in enzyme targets.
ID
TOPIa
NMTb
Cypc
OASSd
51
52
53
54
55
56
57
-98.89
-103.97
-103.79
-115.74
-109.16
-104.11
-107.26
-87.06
-87.05
-87.04
-97.42
-99.07
-96.89
-89.87
-77.36
-77.43
-77.43
-72.70
-72.40
-70.99
-70.80
-102.67
-102.52
-102.53
-103.04
-114.15
-96.16
-99.96
aTopoisomerase I; bN-myristoyltransferase; ccyclophilin; dO-acetylserine sulfhydrylase.
Table 6. Activity of cyclic imides with the best performance in a docking study with each enzyme. pc indicates the probability of
activity from the Combined approach and Lipinski filters.
Combined Probabilities (pc)
Lipinski Filters*
DL
ID
TOPIa
NMTb
Cypc
OASSd
clogP
DS
516
525
531
524
532
52
0.60
0.60
0.61
0.58
0.59
0.57
0.55
0.56
0.62
0.59
0.59
0.56
0.57
0.58
0.55
0.53
0.68
0.66
0.63
0.62
0.60
0.63
0.60
0.60
0.63
0.58
0.55
0.58
0.56
0.54
0.50
0.51
3.71
5.28
5.28
5.16
5.32
4.44
4.32
4.90
-1.74
1.80
0.0
0.33
0.42
0.34
0.49
0.49
0.21
0.23
0.54
4.83
2.88
-7.23
-5.38
2.35
53
522
* clogP (related to lipophilicity); DL (Drug-likeness); and DS (Drug-score). Values calculated using the OSIRIS Property Explorer program available at www.organic-
chemistry.org/prog/peo/
aTopoisomerase I; bN-myristoyltransferase; ccyclophilin; dO-acetylserine sulfhydrylase.
A computational chemistry multitarget model to predict
the results of experimental tests for Leishmania with signifi-
cant success has been reported in the literature [33], so to
finish the study, the combined probabilities of both Ligand
approach and Structure approach were calculated by the
formula: pc = ps + (1 + TN) × p/2 + TN; where pc = Com-
bined approach probability, ps = Structure approach prob-
ability, p = Ligand approach probability, TN = True Nega-
against all enzymes, with probabilities above 50% (Table 6).
The compounds 516, 525, 531, 524, 532, 52, 53 and 522 presents
activity against the four enzymes. Analyzing the RF model
(Table 4) it is possible to observe that the eight compounds
were also classified as active by the model. Fig. (2) summa-
rizes the virtual-screening methodology.
Fig. (3) shows hydrogen-bond interactions between cy-
clic imide 516’s best position in Topoisomerase I (TOPI), N-
myristoyltransferase (NMT), cyclophilin (Cyp) and O-
acetylserine sulfhydrylase (OASS). It is possible to observe
that there is no pattern of interaction between the molecule
and the active sites of the enzymes.
tive probability and ps = Ei/Emin if Ei < Eligand
.
We used our in silico results of the cyclic imides to select
structures that presented lower energy binding (eight com-
pounds) from each enzyme (Table 5). After this step, looking
for multitarget compounds, we selected imides with activity

46 Medicinal Chemistry, 2020, Vol. 16, No. 1
Luis et al.
33 cyclic imides
CHEMBL databank - Inhibitory activity against promastigote Leishmania amazonensis
and
Inhibitory activity against promastigote Leishmania donovani
PDB-Databank IDs:
2B9S, 5A27, 3EOV, 3T4P
Remove duplicate data
3D Generation
CHEMAXON
Generate descriptors
(VOLSURF + )
KNIME
Docking
(Molegro 6.0.1)
Train
Test
Partition
Random Forest
Cross-Validation
Stratified sampling
train/test (80%/20%)
Cyclic imides 10% lowest
MolDock energy
Selected Model
Consensus Analysis
Cyclic imides classified as actives
Selected Compounds
Fig. (2). Virtual-screening methodology used in this study. Solid black lines represent the set of 818 or 722 compounds used to generate the
RF model and to validate it (solid gray line-external test set). Black round dotted lines represent the cyclic imides. Black dash-dot line repre-
sents both datasets (818 or 722 compounds and 33 cyclic imides). The black dash line represents the six enzyme structures from the PDB
databank (2B9S, 5A27, 3EOV, 3T4P). The dash-dot border delimits the process performed in the KNIME software.
Scheme 1. The general synthetic procedure of the target compounds.
3.2. Chemistry
rangement [32]. The final step is the reaction of 11,12-
methylenedioxy-5-methyl-3,4,5,6-tetrahydronaphthalene-2,
15-dicarboxylic anhydride (4) with different aromatic amines
in acetic acid under reflux conditions. The mechanism of the
reaction occurs simultaneously, starting with the attack of
the nucleophile of the amino group to the carbonyl of the
imidic ring because the carbon of the double bond is polar-
ized, having partial positive charge. In this way, the nucleo-
phile can enter, after which the ring ruptures through the exit
of a molecule of H2O (favored by the high temperatures),
forming the corresponding cyclic imides (Scheme 1).
Seven molecules were synthesized. The first intermediate
prepared was isosafrole. It was produced from safrole in a
basic reaction medium through an isomerization. Reflux was
used with KOH/n-butanol, which is a simple and cheap reac-
tion and provides products with good yields [34]. The second
step was the reaction of isosafrole with maleic anhydride
using xylene as the solvent at 100 °C. This reaction is a cy-
clization (4 + 2), similar to a Diels-Alder reaction, the reac-
tion cannot be considered purely Diels-Alder, which happens
in a single phase and this undergoes an additional rear-

Synthesis of New Cyclic Imides Derived from Safrole, Structure- and Ligand-based
Medicinal Chemistry, 2020, Vol. 16, No. 1 47
Trp 15
Gly 316(A)
Asn 317(A)
Phe 14
His 12
Val 315(A)
NMT
TOPI
Ser 354(A)
Gln 133(A)
Gly 81(A)
Ser 79(A)
Thr 83(A)
Gln 86(A)
CyP
OASS
Fig. (3). Lowest-energy dock positions and energies of imide 516 with TOPI (-135.65), NMT (-110.06), CyP (-105.80) and OASS (-112.61).
Dashed lines represent hydrogen-bond interactions. (A higher resolution / colour version of this figure is available in the electronic copy of
the article).
3.2.1. Structural Characterization of Cyclic Imides
136.7, 128.6 and 153.3 ppm of the (C-7), (C-8), (C-11), (C-
12), (C-16), (C-18) and (C-19) carbons, respectively.
As a parameter for the characterization of cyclic imides,
compound 51 was used as an example for the discussion of
the main signals observed in the molecules under study. As-
signments of the hydrogens and carbons performed for com-
pound 51 were based on data obtained from five NMR ex-
The HETeronuclear CORrelation (HETCOR-1JC-H) two-
dimensional (2-D) spectra analysis allowed correlating the
13C nuclei with the 1H directly attached to them (coupled) at
δ 3.52 (H-3) with 43.9 (C-3); δ 4.27 (H-4) with 43.4 (C-4); δ
2.26 (H-5) with 29.9 (C-5); δ 2.61 (H-6) with 34.5 (C-6); δ
6.74 (H-9) with 108.6 (C-9); δ 7.09 (H-10) with 109.3 (C-
10); δ 5.98 (H-13) with 100.8 (C-13); δ 1.13 (H-14) with
16.5 (C-14); δ 8.17 (H-17) with 124.7 (C-17); δ 7.78-7.70
(H-20) with 119.1 (C-20) and δ 7.78-7.70 (H-21) with 135.0
(C-21) ppm (Table 7).
1
periments, namely, H NMR, 13C-APT NMR, COSY, HET-
COR and HMBC. Table 7 summarizes the assignments made
to each hydrogen and carbon of the molecule. The attribu-
tions of the hydrogens and carbons follow the numbering
shown in Fig. (4). In the analysis of the 13C-APT NMR spec-
tra at 50 MHz of 51, the presence of 20 signals was observed,
of which nine down signals were associated with hydrogen-
ated carbons, one of which was assigned to trihydrogenated
carbon at δ 16.54 ppm of (C-3), (C-4) and (C-5), and five of
the aromatic carbons at δ 108.6, 109.3, 124.7, 119.1 and
135.0 ppm of carbons (C-9), (C-10), (C-17), (C-20) and (C-
21), respectively. The remaining 11 signals, all up, two cor-
responded to aliphatic dihydrogenated carbons at δ 34.5 and
108.8 ppm of (C-6) and (C-13), respectively. Nine
unhydrogenated carbons of the sp2 type, two of carbonyl
carbons of (C-2) and (C-15) at δ 176.5 and 175.9 ppm
respectively and seven quaternary carbons at δ 130.19,
122.1, 145.7, 146.3, 136.7, 128.6 and 153.3 ppm of the (C-
Analysis of the 2-D spectroscopy (COSY-3JH-H and
1
1
-4JH-H) spectra allowed to correlate the H nuclei with H
remote three and four bonds, respectively. We start our stud-
ies from the hydrogens of the methyl group, which is a good
starting group. It appears as an intense doublet of hydrogens
(H-14) at δ 1.13 ppm (J = 7.2 Hz), which only couples with
hydrogen (H-5) at δ 2.26 ppm. This coupling confirms that
the CH3 group is terminal. The hydrogen of the (H-5) group
at δ 2.26 ppm is also coupled with the hydrogen (H-6) at δ
2.61 ppm and (H-3) at δ 3.52 ppm, which, in turn, couples
with the (H-4) at δ 4.27 ppm. Hydrogen (H-6) at δ 2.61 ppm

48 Medicinal Chemistry, 2020, Vol. 16, No. 1
Luis et al.
Table 7. 1H NMR (200 MHz) and 13C (50 MHz) data on (DMSO-d6) of 51 (Scheme 1). The atom labeled as 1 is nitrogen. The
chemical shifts are ppm.
HETCOR/APT
Carbon
δ (ppm)
COSY
HMBC
c, e
d
δ (2J and 3JH-H
)
δ (2J and 3JC-H
)
δ (13C)a
δ (1H)b, e
2
3
176.5
-
-
-
3.52 (dd, J = 8.9, 5.7 Hz,
4.27 (d, J = 9.0 Hz, H-4) and 2.26 (m,
16.5 (C-14); 29.9 (C-5); 176.5 (C-2)
and 175.9 (C-15)
43.9
43.4
1H)
H-5)
43.9 (C-3); 122.1 (C-8); 176.5 (C-2)
and 175.9 (C-15)
4
5
6
4.27 (d, J = 9.0 Hz, 1H)
2.26 (m, 1H)
3.52 (dd, J = 8.9, 5.7 Hz, 1H)
3.52 (dd, J = 8.9, 5.7 Hz, H-3), 2.61
(m, H-6) and 1.13 (d, J = 7.2 Hz, H-
14)
29.9
34.5
34.5 (C-6); 43. 9(C-3) and 16.5 (C-14)
43.9 (C-3); 29.9 (C-5); 108.6 (C-9) and
130.2 (C-7)
2.61 (m, 2H)
2.26 (m, H-5)
7
8
130.2
122.1
-
-
-
-
-
-
34.5 (C-6); 122.1 (C-8); 145.7 (C-11)
and 146.3 (C-12)
9
108.6
109.3
6.74 (s, 1H)
7.09 (s, 1H)
-
-
43.4 (C-4); 130.2 (C-7); 145.7 (C-11)
and 146.3 (C-12)
10
11
12
13
14
15
16
17
18
19
20
145.7
146.3
100.8
16.5
-
-
-
-
-
-
5.98 (d, J = 15,8 Hz, 2H)
-
145.7 (C-11) and 146.3 (C-12)
1.13 (d, J = 7.2 Hz, 3H)
2.26 (m, H-5)
43.9 (C-3) and 29.9 (C-5)
175.9
136.7
124.7
128.6
153.9
119.1
-
-
-
-
-
-
8.17 (d, J = 2.0 Hz, 1H)
7.78-7.70 (m, H-21)
153.9 (C-19) and 135.0 (C-21)
-
-
-
-
-
-
7.78-7.70 (m, 1H)
7.78-7.70 (m, H-21)
128.6 (C-18) and 135.0 (C-21)
7.78-7.70 (m, H-21) and 8.17 (d, J =
21
135.0
7.78-7.70 (m, 1H)
124.7 (C-17) and 153.9 (C-19)
2.0 Hz, H-17)
aValues deduced by 13C-APT NMR spectra; bValues obtained from 2-D heteronuclear correlations through one bond (1JC-H) HETCOR; cValues obtained from 2-D homonuclear
4
correlations through a (3JH-H and JH-H) COSY; dValues obtained from 2-D correlations through (2JC-H) and (3JC-H) HMBC. eMultiplicity of signals for 1H NMR: singlet (s); doublet
(d); double doublet (dd); quartet (q); septet (sept) and multiplet (m).
and (H-4) at δ 4.27 ppm, were coupled once, thus showing
that these groups are also terminal (Table 7). In the same
spectra, we can observe three important correlations in the
aromatic region. The COSY spectra show a doublet at δ 8.17
ppm (J = 2.0 Hz) of (H-17), which is long-distance meta-
coupled (4JH-H) with hydrogen (H-21) at δ 7.78-7.70 ppm,
which in turn, three ortho bonds (3JH-H) are coupled ortho
to hydrogen (H-21) at δ 7.78-7.70 ppm. The use of (COSY-
3JH-H and -4JH-H) was undoubtedly a technique of great im-
portance in the attribution of the chemical displacements of
hydrogens (H-3), (H-4), (H-6) and (H-14) of aliphatic and
(H-17), (H-20) and (H-21) of aromatics according to Fig. (4).
According to the 2-D Heteronuclear Multiple Bond Co-
herence spectra (HMBC-2J and -3JC-H) it was possible to
1
unequivocally assign the couplings between 13C and H dis-
tant two and three bonds from the hydrogen coupling (H-14)
of the methyl group at δ 1.13 ppm with the carbons (C-5)
and (C-3) at δ 29.9 and 43.9 ppm respectively, in turn, the

Synthesis of New Cyclic Imides Derived from Safrole, Structure- and Ligand-based
Medicinal Chemistry, 2020, Vol. 16, No. 1 49
Fig. (4). (A) 1H-1H COSY correlations of compound 51. (B) HMBC-2J and -3JC-H correlations of compound 51.
hydrogen (H-5) at δ 2.62 ppm next to the methyl group made
two correlations with the carbons (C-6) and (C-3) at δ 34.5
and 43.9 ppm, respectively. Two important correlations were
attributed to hydrogen atoms isolated from hydrogen (H-13)
at δ 100.8 ppm with carbons (C-11) and (C-12) at δ 145.7
and 146.3 ppm, respectively. Other important correlations
was the strain available in our laboratories and the docking
study for L. donovani, because it was the species with
enzymes characterized and registered in PDB. Each study
provided useful information for future works. With the
material available in our laboratories we synthesized seven
imides and tested them against L. amazonensis promastigotes
and all were inactive (pIC50 <5) as predicted in the Ligand
based approach (LB). The Structure based approach (SB) for
the enzymes CyP, NMT, OASS and TOPI of L. donovani,
together with the LB, allowed the visualization of some
important characteristics of the molecules that were
predicted as active in our model. Molecules 516, 52 and 53
were the most likely to be active in the LB for L. donovani
(82, 81 and 81%, respectively). These three molecules
present nitro groups in the meta position of the aromatic ring
acting as hydrogen bonding acceptors, but without
interaction pattern with specific amino acids, as can be
observed in Fig. (3). Molecules 525, 531, 524, 532 and 522, with
LB probabilities of activities of 75, 71, 71, 53, and 68%,
respectively, are molecules which, instead of phenyl
substituents, have benzylic groups substituted on the imidic
nitrogen with lipophilic groups in meta or para position.
2
3
were attributed from J and JC-H couplings at: δ 3.52 (H-3)
with 176.5 (C-2); 29.9 (C-5); 16.5 (C-14) and 175.9 (C-15)
ppm; δ 4.27 (H-4) with 176.5 (C-2); 43.9 (C-3); 122.1 (C-8)
and 175.9 (C-15) ppm; δ 2.26 (H-6) with 43.9 (C-3); 29.9
(C-5); 130.2 (C-7) and 108.6 (C-9) ppm. The aromatic hy-
drogens had their attributions confirmed from the long-
distance correlations (HMBC-2J and -3JC-H) at δ 6.74 (H-9)
with 34.5 (C-6); 122.1 (C-8); 145.7 (C-11) and 146.3 (C-12)
ppm; δ 6.74 (H-10) with 43.4 (C-4); 130.2 (C-7); 145.7 (C-
11) and 146.3 (C-12) ppm; δ 8.17 (H-17) with 153.9 (C-19)
and 135.0 (C-21) ppm; δ 7.78-7.70 (H-20) with 136.7 (C-
16); (C-18) and 135.0 (C-21) ppm and finally hydrogen at δ
6.74 (H-21) with 124.7 (C-17) and 153.9 (C-19) ppm (Fig.
4). The results indicated that the combined use of the one-
1
dimensional (1-D) H and 13C-APT and two-dimensional (2-
D) COSY, HETCOR and HMBC NMR techniques allowed
proposing the basic skeleton of the hydrogens and carbons of
51 (Table 1) and these results were important for the assign-
ment of signals from molecules 52-7 found in Section 2.
To try to rationalize the results we used Lipinski filters
through the OSIRIS Property Explorer program to analyze
the lipophilicity of molecules (clog P), Drug-likeness (DL)
that evaluates the comparison of the investigated compounds
through fragments and/or physical properties similar to those
of the most known drugs, and Drug-score (DS) that com-
bines the records of drug-likeness, lipophilicity, solubility,
molecular mass and toxicity risks into a single numerical
value, which varies from 0.0 to 1.0 and can be used to pre-
dict the global potential of one compound as a new drug
candidate. It was observed that compounds 516, 52 and 53,
because they have nitro groups with great toxic potential,
were the compounds with lower DL and DS values (Table
6). Compound 524 which has a benzyl group with a chlorine
atom substituted in para-position was the one with the
highest values of DL and DS (Table 6).
To verify the effectiveness of the virtual-screening
method, we performed in vitro tests with the seven mole-
cules synthesized (51-7) against promastigote forms of L.
amazonensis (IFLA/BR/1967/PH8). The results showed
pIC50 values 3.56, 3.06, 3.01, 2.98, 3.43, 3.46, 2.98, respec-
tively. In the ligand-based approach, these seven molecules
were predicted to be inactive (probabilities: 61, 53, 57, 74,
68, 63 and 65%, respectively) (Table 3), so the results of
biological activity corroborate with those of in silico analy-
sis, considering that our model was better at predicting inac-
tive molecules (Table 1). It is important to highlight that this
was a preliminary work, carried out to obtain candidates for
drugs against L. amazonensis and L. donovani. We
performed the in vitro tests against L. amazonensis, since it

50 Medicinal Chemistry, 2020, Vol. 16, No. 1
Luis et al.
Then, based on the findings of this study, we are
proposing synthesize new molecules with benzylic
substituents on imidic nitrogen with donor and acceptor of
hydrogen bonds with meta-substituents and lipophilic groups
in the para position. These new compounds will have great
potential to act on multiple targets in parasites of the genus
Leishmania.
REFERENCES
[1]
Alvar, J.; Vélez, I.D.; Bern, C.; Herrero, M.; Desjeux, P.; Cano, J.;
Jannin, J.; den Boer, M.; Team, W.L.C. WHO leishmaniasis con-
trol team. Leishmaniasis worldwide and global estimates of its in-
cidence. PLoS One, 2012, 7(5), e35671.ꢀ
[http://dx.doi.org/10.1371/journal.pone.0035671]
[PMID: 22693548]
[2]
Rodrigues, K.A.F.; Dias, C.N.S.; Néris, P.L.N.; Rocha, J.C.; Scotti,
M.T.; Scotti, L.; Mascarenhas, S.R.; Veras, R.C.; de Medeiros,
I.A.; Keesen, Tde.S.; de Oliveira, T.B.; de Lima, M.C.; Balliano,
T.L.; de Aquino, T.M.; de Moura, R.O.; Mendonça Junior, F.J.; de
Oliveira, M.R. 2-Amino-thiophene derivatives present antileishma-
nial activity mediated by apoptosis and immunomodulation in vi-
tro. Eur. J. Med. Chem., 2015, 106, 1-14.ꢀ
CONCLUSION
In this study, we have conducted a comparative ligand-
and structure-based approach using Molegro Virtual Dock-
ing and machine learning RF to determine the anti-
Leishmanial potential of thirty-three cyclic imides set. In
silico study allowed us to suggest that the cyclic imides 516,
525, 531, 524, 532, 52, 53 and 522 can be tested as potential mul-
titarget molecules for leishmanial treatment, presenting ac-
tivity against four strategic enzymes for treatment with a
probability of activity above 60%. So, to evaluate the
efficiency of the computational method used, we have syn-
[http://dx.doi.org/10.1016/j.ejmech.2015.10.011]
[PMID: 26513640]
[3]
[4]
Bonano, V.I.; Yokoyama-Yasunaka, J.K.; Miguel, D.C.; Jones,
S.A.; Dodge, J.A.; Uliana, S.R. Discovery of synthetic Leishmania
inhibitors by screening of a 2-arylbenzothiophene library. Chem.
Biol. Drug Des., 2014, 83(3), 289-296.ꢀ
[http://dx.doi.org/10.1111/cbdd.12239] [PMID: 24119198]
Herrera Acevedo, C.; Scotti, L.; Feitosa Alves, M.; Formiga Melo
Diniz, M.F.; Scotti, M.T. Computer-aided drug design using ses-
quiterpene lactones as sources of new structures with potential ac-
tivity against infectious neglected diseases. Molecules, 2017, 22(1),
79.ꢀ
1
thesized and characterized seven cyclic imides by IR, H
NMR, 13C-APT NMR, COSY, HETCOR and HMBC. The
compounds tested against promastigote forms of L. ama-
zonensis presented pIC50 values less than 4.7, showing that
our method was efficient in predicting true negative
molecules.
[http://dx.doi.org/10.3390/molecules22010079] [PMID: 28054952]
Thompson, A.M.; O’Connor, P.D.; Marshall, A.J.; Yardley, V.;
Maes, L.; Gupta, S.; Launay, D.; Braillard, S.; Chatelain, E.;
Franzblau, S.G.; Wan, B.; Wang, Y.; Ma, Z.; Cooper, C.B.; Denny,
W.A. 7-substituted 2-nitro-5,6-dihydroimidazo[2,1-b][1,3]oxa-
zines: Novel antitubercular agents lead to a new preclinical candi-
date for visceral leishmaniasis. J. Med. Chem., 2017, 60(10), 4212-
4233.ꢀ
[5]
ETHICS APPROVAL AND CONSENT TO PARTICI-
PATE
[http://dx.doi.org/10.1021/acs.jmedchem.7b00034]
Not applicable.
[PMID: 28459575]
[6]
[7]
Ogungbe, I.V.; Erwin, W.R.; Setzer, W.N. Antileishmanial phyto-
chemical phenolics: molecular docking to potential protein targets.
J. Mol. Graph. Model., 2014, 48, 105-117.ꢀ
HUMAN AND ANIMAL RIGHTS
No Animals/Humans were used for studies that are base
of this research.
[http://dx.doi.org/10.1016/j.jmgm.2013.12.010] [PMID: 24463105]
Hermoso, A.; Jiménez, I.A.; Mamani, Z.A.; Bazzocchi, I.L.; Pi-
ñero, J.E.; Ravelo, A.G.; Valladares, B. Antileishmanial activities
of dihydrochalcones from piper elongatum and synthetic related
compounds. Structural requirements for activity. Bioorg. Med.
Chem., 2003, 11(18), 3975-3980.ꢀ
CONSENT FOR PUBLICATION
Not applicable.
[http://dx.doi.org/10.1016/S0968-0896(03)00406-1]
[PMID: 12927858]
AVAILABILITY OF DATA AND MATERIALS
[8]
[9]
Lill, M.A.; Danielson, M.L. Computer-aided drug design platform
using PyMOL. J. Comput. Aided Mol. Des., 2011, 25(1), 13-19.ꢀ
[http://dx.doi.org/10.1007/s10822-010-9395-8] [PMID: 21053052]
Lorenzo, V.P.; Lúcio, A.S.; Scotti, L.; Tavares, J.F.; Filho, J.M.;
Lima, T.K.; Rocha, J.D.; Scotti, M.T. Structure- and ligand-based
approaches to evaluate aporphynic alkaloids from annonaceae as
multi-target agent against Leishmania donovani. Curr. Pharm.
Des., 2016, 22(34), 5196-5203.ꢀ
Elemental analysis for compounds 51-7, 1H and 13C
NMR spectra for compound 51, as well as the smiles of the
compounds used in the databases is available from the corre-
sponding author, Luis JAD, upon reasonable request.
FUNDING
[http://dx.doi.org/10.2174/1381612822666160513144853] [PMID:
27174814]
This work was supported by Brazilian National Council
for Scientific and Technological Development (CNPq),
award Number 310919/2016-9 and 431254/2018-4.
[10]
Lorenzo, V.P.; Barbosa-Filho, J.M.; Scotti, L.; Scotti, M.T. Com-
bined structure- and ligand-based virtual screening to evaluate
caulerpin analogs with potential inhibitory activity against mono-
amine oxidase B. Rev. Bras. Farmacogn., 2015, 25(6), 690-697.ꢀ
[http://dx.doi.org/10.1016/j.bjp.2015.08.005]
CONFLICT OF INTEREST
[11]
[12]
Hargreaves, M.K.; Pritchard, J.; Dave, H. Cyclic carboxylic mono-
imides. Chem. Rev., 1970, 70(4), 439-469.ꢀ
The authors declare no conflict of interest, financial or
otherwise.
[http://dx.doi.org/10.1021/cr60266a001]
Cechinel-Filho, V. Principais avanços e perspectivas na área de
ACKNOWLEDGEMENTS
produtos
naturais
ativos:
estudos
desenvolvidos
no
NIQFAR/UNIVALI. Quim. Nova, 2000, 23(5), 680-685.ꢀ
Declared none.
[http://dx.doi.org/10.1590/S0100-40422000000500017]
[13]
Cechinel Filho, V.; Pinheiro, T.; Nunes, R.J.; Yunes, R.A.; Cruz,
A.B.; Moretto, E. Antibacterial activity of N-phenylmaleimides, N-
phenylsuccinimides and related compounds. Structure-activity rela-
tionships. Farmaco, 1994, 49(10), 675-677.ꢀ
SUPPLEMENTARY MATERIAL
Supplementary material is available on the publisher’s
web site along with the published article.
[PMID: 7826477]

Synthesis of New Cyclic Imides Derived from Safrole, Structure- and Ligand-based
Medicinal Chemistry, 2020, Vol. 16, No. 1 51
[25]
Hall, M.; Frank, E.; Holmes, G.; Pfahringer, B.; Reutemann, P.;
Witten, I.H. The WEKA data mining software: an update. Explora-
tions newsletter, 2009, 11(1), 10-18.
[14]
[15]
[16]
Yunes, J.A.; Cardoso, A.A.; Yunes, R.A.; Correa, R.; de Campos-
Buzzi, F.; Cechinel-Filho, V.Z. Antiproliferative effects of a series
of cyclic imides on primary endothelial cells and a leukemia cell
line. Naturforsch. C, 2008, 63(9-10), 675-680.ꢀ
[http://dx.doi.org/10.1515/znc-2008-9-1011] [PMID: 19040106]
Zawadowski, T.; Kossakowski, J.; Rump, S.; Jakowicz, I.; Płaźnik,
A. Synthesis and anxiolytic activity of N-substituted cyclic imides
N-[4-[(4-aryl)-1-piperazinyl]alkyl]-5,7-dioxabicyclo[2.2.2]octane-
2, 3-dicarboximide. Acta Pol. Pharm., 1995, 52(1), 43-46.ꢀ
[PMID: 8960237]
Alaa, A-M. Novel and versatile methodology for synthesis of cy-
clic imides and evaluation of their cytotoxic, DNA binding, apop-
totic inducing activities and molecular modeling study. Eur. J.
Med. Chem., 2007, 42(5), 614-626.ꢀ
[http://dx.doi.org/10.1016/j.ejmech.2006.12.003]
[PMID: 17234303]
Mesomo, M.C.; Corazza, M.L.; Ndiaye, P.M.; Dalla Santa, O.R.;
Cardozo, L.; Scheer, A.R.P. Supercritical CO2 extracts and essen-
tial oil of ginger (Zingiber officinale R.): Chemical composition
and antibacterial activity. J. Supercrit. Fluids, 2013, 80, 44-49.ꢀ
[http://dx.doi.org/10.1016/j.supflu.2013.03.031]
Povh, N.P.; Marques, M.O.; Meireles, M.A.A. Supercritical CO2
extraction of essential oil and oleoresin from chamomile (Chamo-
milla recutita [L.] Rauschert). J. Supercrit. Fluids, 2001, 21(3),
245-256.ꢀ
[http://dx.doi.org/10.1016/S0896-8446(01)00096-1]
Imre, G.; Veress, G.; Volford, A.; Farkas, Ö. Molecules from the
Minkowski space: an approach to building 3D molecular structures.
J. Mol. Struct. THEOCHEM, 2003, 666, 51-59.ꢀ
[http://dx.doi.org/10.1016/j.theochem.2003.08.013]
Cruciani, G.; Crivori, P.; Carrupt, P-A.; Testa, B. Molecular fields
in quantitative structure–permeation relationships: the VolSurf ap-
proach. J. Mol. Struct. THEOCHEM, 2000, 503(1), 17-30.ꢀ
[http://dx.doi.org/10.1016/S0166-1280(99)00360-7]
Scotti, L.; Ishiki, H.; Mendonça Júnior, F.J.; Da Silva, M.S.; Scotti,
M.T. In-silico analyses of natural products on leishmania enzyme
targets. Mini Rev. Med. Chem., 2015, 15(3), 253-269.ꢀ
[http://dx.doi.org/10.2174/138955751503150312141854] [PMID:
25769973]
[https://doi.org/10.1145/1656274.1656278]
[26]
Hanley, J.A.; McNeil, B.J. The meaning and use of the area under a
receiver operating characteristic (ROC) curve. Radiology, 1982,
143(1), 29-36.ꢀ
[http://dx.doi.org/10.1148/radiology.143.1.7063747]
[PMID: 7063747]
Davies, D.R.; Mushtaq, A.; Interthal, H.; Champoux, J.J.; Hol,
W.G. The structure of the transition state of the heterodimeric
topoisomerase I of Leishmania donovani as a vanadate complex
with nicked DNA. J. Mol. Biol., 2006, 357(4), 1202-1210.ꢀ
[http://dx.doi.org/10.1016/j.jmb.2006.01.022] [PMID: 16487540]
Rackham, M.D.; Yu, Z.; Brannigan, J.A.; Heal, W.P.; Paape, D.;
Barker, K.V.; Wilkinson, A.J.; Smith, D.F.; Leatherbarrow, R.J.;
Tate, E.W. Discovery of high affinity inhibitors of Leishmania
donovani N-myristoyltransferase. MedChemComm, 2015, 6(10),
1761-1766.ꢀ
[http://dx.doi.org/10.1039/C5MD00241A] [PMID: 26962429]
Venugopal, V.; Datta, A.K.; Bhattacharyya, D.; Dasgupta, D.;
Banerjee, R. Structure of cyclophilin from Leishmania donovani
bound to cyclosporin at 2.6 A resolution: correlation between struc-
ture and thermodynamic data. Acta Crystallogr. D Biol. Crystal-
logr., 2009, 65(Pt 11), 1187-1195.ꢀ
[http://dx.doi.org/10.1107/S0907444909034234]
[PMID: 19923714]
Raj, I.; Kumar, S.; Gourinath, S. The narrow active-site cleft of O-
acetylserine sulfhydrylase from Leishmania donovani allows com-
plex formation with serine acetyltransferases with a range of C-
terminal sequences. Acta Crystallogr. D Biol. Crystallogr., 2012,
68(Pt 8), 909-919.ꢀ
[http://dx.doi.org/10.1107/S0907444912016459]
[PMID: 22868756]
Thomsen, R.; Christensen, M.H. MolDock: a new technique for
high-accuracy molecular docking. J. Med. Chem., 2006, 49(11),
3315-3321.ꢀ
[http://dx.doi.org/10.1021/jm051197e] [PMID: 16722650]
Hudson, B.; Robinson, R. Addition of maleic anhydride and ethyl
maleate to substituted styrenes. J. Chem. Soc., 1941, 715-722.ꢀ
[http://dx.doi.org/10.1039/jr9410000715]
García, I.; Fall, Y.; Gómez, G.; González-Díaz, H. First computa-
tional chemistry multi-target model for anti-Alzheimer, anti-
parasitic, anti-fungi, and anti-bacterial activity of GSK-3 inhibitors
in vitro, in vivo, and in different cellular lines. Mol. Divers., 2011,
15(2), 561-567.ꢀ
[27]
[28]
[17]
[18]
[29]
[30]
[31]
[19]
[20]
[21]
[32]
[33]
[22]
Scotti, L.; Scotti, M.T. Computer aided drug design studies in the
discovery of secondary metabolites targeted against age-related
neurodegenerative diseases. Curr. Top. Med. Chem., 2015, 15(21),
2239-2252.ꢀ
[http://dx.doi.org/10.2174/1568026615666150610143510] [PMID:
26059353]
[http://dx.doi.org/10.1007/s11030-010-9280-3] [PMID: 20931280]
Barreiro, E.J.; Lima, M.E. The synthesis and anti-inflammatory
properties of a new sulindac analogue synthesized from natural sa-
frole. J. Pharm. Sci., 1992, 81(12), 1219-1222.ꢀ
[23]
[24]
Berthold, M.R.C.N.; Dill, F. Data analysis, machine learning and
applications. Springer: Berlin, 2007; pp. 36-79.
[34]
Breiman, L. Random forests. Mach. Learn., 2001, 45(1), 5-32.ꢀ
[http://dx.doi.org/10.1023/A:1010933404324]
[http://dx.doi.org/10.1002/jps.2600811219] [PMID: 1491344]