Glucosyl-1,2,3-triazoles derived from eugenol and analogues: Synthesis, anti-Candida activity, and molecular modeling studies in CYP-51

Article abstract of DOI:10.1111/cbdd.13948

This work describes the synthesis, anti-Candida, and molecular modeling studies of eighteen new glucosyl-1,2,3-triazoles derived from eugenol and correlated phenols. The new compounds were characterized by combined Fourier Transform Infrared, ¹H and ¹³C nuclear magnetic resonance and spectroscopy of high-resolution mass spectrometry. The synthesized compounds did not show significant cytotoxicity against healthy fibroblast human cells (MCR-5) providing interesting selectivity indexes (SI) to active compounds. Considering the antifungal activity, nine compounds showed anti-Candida potential and the peracetylated triazoles 17 and 18 were the most promising ones. Eugenol derivative 17 was active against three species of Candida at 26.1–52.1?μM. This compound was four times more potent than fluconazole against Candida krusei and less toxic (SI?>?6.6) against the MCR-5 cells than fluconazole (SI?>?3.3) considering this strain. Dihydroeugenol derivative 18?showed similar activity to 17 and was four times more potent and less toxic than fluconazole against C.?krusei. The deacetylated glucosides and non-glucosylated corresponding derivatives did not show considerable antifungal action, suggesting that the acetyl groups are essential for their anti-Candida activity. Molecular docking coupled with molecular dynamics showed that 14α-lanosterol demethylase is a feasible molecular target, since 17 and 18 could bind to this enzyme once deacetylated in vivo, thereby acting as prodrugs. Also, these studies demonstrated the importance of hydrophobic substituents at the phenyl ring.

Full text of DOI:10.1111/cbdd.13948

Received: 18 June 2021
Revised: 3 August 2021
Accepted: 30 August 2021
|
|
DOI: 10.1111/cbdd.13948
R E S E A R C H A R T I C L E
Glucosyl-1,2,3-triazoles derived from eugenol and
analogues: Synthesis, anti-Candida activity, and molecular
modeling studies in CYP-51
Lorena Severiano de Magalhães1
Adriana Cotta Cardoso Reis1
|
|
Izadora Amaral Nakao1
Vinícius Augusto Campos Péret1
|
|
Rúbia Castro Fernandes Melo Reis1
Naiara Chaves Silva2
|
|
Amanda Latércia Tranches Dias2
Diogo Teixeira Carvalho3
|
|
Danielle Ferreira Dias4
Geraldo Célio Brandão1
Saulo Fehelberg Pinto Braga1
|
|
|
Thiago Belarmino de Souza1
1School of Pharmacy, Federal
University of Ouro Preto, Ouro Preto,
Brazil
2Institute of Biomedical Sciences,
Federal University of Alfenas, Alfenas,
Brazil
3Pharmaceutical Sciences University,
Federal University of Alfenas, Alfenas,
Brazil
4Institute of Chemistry, Federal
Abstract
This work describes the synthesis, anti-Candida, and molecular modeling stud-
ies of eighteen new glucosyl-1,2,3-triazoles derived from eugenol and corre-
lated phenols. The new compounds were characterized by combined Fourier
Transform Infrared, H and 13C nuclear magnetic resonance and spectroscopy
1
of high-resolution mass spectrometry. The synthesized compounds did not show
significant cytotoxicity against healthy fibroblast human cells (MCR-5) providing
interesting selectivity indexes (SI) to active compounds. Considering the antifun-
gal activity, nine compounds showed anti-Candida potential and the peracety-
lated triazoles 17 and 18 were the most promising ones. Eugenol derivative 17
was active against three species of Candida at 26.1–52.1 μM. This compound was
four times more potent than fluconazole against Candida krusei and less toxic
(SI > 6.6) against the MCR-5 cells than fluconazole (SI > 3.3) considering this
strain. Dihydroeugenol derivative 18 showed similar activity to 17 and was four
times more potent and less toxic than fluconazole against C. krusei. The deacety-
lated glucosides and non-glucosylated corresponding derivatives did not show
considerable antifungal action, suggesting that the acetyl groups are essential for
their anti-Candida activity. Molecular docking coupled with molecular dynam-
ics showed that 14α-lanosterol demethylase is a feasible molecular target, since
17 and 18 could bind to this enzyme once deacetylated in vivo, thereby acting as
prodrugs. Also, these studies demonstrated the importance of hydrophobic sub-
stituents at the phenyl ring.
University of Alfenas, Alfenas, Brazil
Correspondence
Thiago Belarmino de Souza, School of
Pharmacy, Federal University of Ouro
Preto, Ouro Preto, Brazil.
Email: thiagobs83@yahoo.com.br
Funding information
Fundação de Amparo à Pesquisa do
Estado de Minas Gerais; Conselho
Nacional de Desenvolvimento
Científico e Tecnológico
K E Y W O R D S
anti-Candida activity, CYP51, eugenol, glycosyl-triazoles, molecular docking, molecular
dynamics
Chem Biol Drug Des. 2021;00:1–11.
wileyonlinelibrary.com/journal/cbdd
© 2021 John Wiley & Sons Ltd.
1
|

2
MAGALHÃES et al.
|
1
INTRODUCTION
bioactive substance can lead to an increase in its biological
activity by creating new points of attachment with a tar-
get of the microorganism. Therefore, we recently reported
the synthesis of eugenol glucosides with antifungal po-
tential: Derivative II was active against Candida glabrata
at 3.4 µM (3.4 times more potent than fluconazole), and
derivative III showed fungistatic (IC50: 18.1 µM) and fun-
gicidal (IC50: 36.2 µM) activities against this same strain
(Souza et al., 2014, 2016).
Based on the anti-Candida activity observed for glyco-
sidic derivatives II and III, we report herein the synthesis
and evaluation of new glucosyl triazoles derivatives of eu-
genol and related substances (dihydroeugenol, guaiacol,
phenol, vanillin, and isoeugenol) with the structural pat-
tern showed in Figure 2. In addition to the glucosylated
compounds (in peracetylated and deacetylated forms), the
respective non-glucosylated derivatives were also synthe-
sized to evaluate the importance of the saccharide moiety
in the antifungal activity. We decided to associate the sac-
charide unit to the phenols through a 1,2,3-triazole ring
from a click reaction, which is already a well-established
reaction by our research group. This heterocycle could
|
Indicators of morbidity and mortality from fungal infec-
tions are increasing at an alarming rate, both in immu-
nocompetent and in immunosuppressed patients. It is
estimated that there are approximately six million fungal
species in the world and about 600 species can cause dis-
eases in humans. These infections affect over one billion
people each year and leading to more than 1.5 million
deaths worldwide annually (Strickland & Shi, 2021).
Candida infection is one of the most common fungal
infections and Candida albicans is the most commonly
recovered from clinical species, followed by Candida
glabrata. Other clinically important species include
C. parapsilosis, C. krusei, C. tropicalis, C. dubliniensis,
C. lusitaniae, and the most recent, C. auris. C. auris is an
emerging multiresistant species that is difficult to diag-
nose and is causing great concern in health centers cur-
rently. The cases associated with C. auris increased by
318% between 2015 and 2018 (Bhattacharya et al., 2020).
In Latin America, these infections are caused mainly by
C. albicans, but other species are also relevant and preva-
lent under certain conditions. For example, vulvovaginal
candidiasis in diabetic patients and patients with hema-
tological diseases is often caused by C. glabrata. Patients
undergoing treatment for myeloid leukemia commonly
present infection by C. tropicalis. Invasive infections by
C. krusei are very relevant in bone marrow transplant pa-
tients. C. parapsilosis causes infections mainly in children
and is the second leading cause of endocarditis (Whaley
et al., 2017).
There are only three classes of available drugs for the
treatment of systemic fungal infections, the polyene am-
photericin B, echinocandins and azoles. Most species of
fungi have already developed resistance to these drugs
and, in addition to resistance problems, these drugs can
lead to adverse effects as nausea and vomiting, besides
some potentially dangerous effects, such as liver and
kidney toxicity, and hepatobiliary disorders (Revie et al.,
2018).
FIGURE 1 Eugenol (I) and glucosyl derivatives (II and III)
with antifungal activity
Lately, our research group has spared no efforts in the
search of new substances with antimicrobial potential
that can increase the therapeutic arsenal for the treatment
of these infections. We have focused mainly on the syn-
thesis of glycosidic derivatives of eugenol (I, Figure 1), a
natural allylphenol that shows several biological activi-
ties, including antifungal. Linking bioactive compounds
to carbohydrate residues (as glucose) could increase the
ability of these compounds to penetrate microorgan-
ism cells through glucose transporters present in fungal
membranes and contribute to modulate important phys-
icochemical properties for drug action, such as solubil-
ity. Furthermore, the insertion of saccharide units in a
FIGURE 2 Structural pattern of glucosylated and non-
glucosylated 1,2,3-triazoles derived from eugenol and analogues

MAGALHÃES et al.
3
|
also act as an isoster of pharmacophore azole rings pres-
ent in different drugs as ketoconazole, miconazole, and
other antifungal agents, and here, we could evaluate the
influence of this group in the mono or disubstituted form
regarding its potential to mimic different azole rings of
these drugs. Moreover, with the proposed compounds, we
were able to have a greater understanding about the im-
portance of carbohydrate and the groups from eugenol to
their antifungal potential against Candida spp.
hydrobromic acid and acetic acid (Guilherme et al., 2019).
The reaction of 15 with sodium azide led to the α-glucosyl
azide 16, which was employed in the next step without
previous purification (Pote et al., 2018). The reaction of
alkynes (7–12) with the glucosyl azide 16 by cooper cata-
lyzed 1,3-dipolar cycloaddition (“click” reaction) resulted
in the peracetylated 1,2,3-triazole derivatives (17–22) with
yields in the range of 51–71%. Copper (I) acts by complex-
ing with the terminal carbon of the alkyne generating a
metal acetylide and the nucleophilic nitrogen of the al-
kylazide attacks the complexed copper, directing the for-
mation of the 1,4-regioisomer (Safavi et al., 2019). The
deacetylated 1,2,3-triazoles (23–28) were obtained with
yields in the range of 30–98% after methanolysis of glu-
cosides 17–22 (Safavi et al., 2019). In order to evaluate the
importance of the saccharide unit in the structure of the
compounds, the respective non-glycosylated triazoles (29–
34) were also synthesized from the reaction of the alkynes
7–12 with sodium azide and acetic acid, thus obtaining
yields in the range of 57–61% (Banert et al., 2016).
2
RESULTS AND DISCUSSION
Chemistry
|
2.1
|
The new triazoles derived from eugenol and analogues
were synthesized according to the synthetic route seen in
Scheme 1.
First, eugenol (1) and the other phenols (2–6) were al-
kylated from the Williamson synthesis, using propargyl
chloride to provide the alkynes 7–12 by the method de-
scribed by Achard et al. (2011). In parallel, D-glucose (13)
was converted to its peracetylated derivative (14) after its
reaction with acetic anhydride, which subsequently pro-
vided the acetobromoglucoside (15) after reaction with
The analysis of infrared spectra of derivatives 17–22
showed typical bands of ester carbonyl and aromatic
C=C stretching near 1740 and 1599–1450 cm−1, respec-
tively. For deacetylated triazoles 23–28, hydroxyl bands
1
near 3300 cm−1 were noticed as expected. In H nuclear
SCHEME 1 Synthesis of glucosylated-1,2,3-triazole derivatives of eugenol and analogues

4
MAGALHÃES et al.
|
magnetic resonance (NMR) spectra of derivatives 17–22,
a singlet was registered near 7.40 ppm corresponding
to the 1,2,3-triazolic ring proton, proving the formation
of this heterocycle. The signals related to the aromatic
hydrogens were recorded between 7.0 and 6.6 ppm, ac-
cording to the substitution pattern of each compound.
The saccharide hydrogens were observed between 6.38
and 3.95 ppm, while the acetyl group hydrogens were
recorded as singlets between 2.91 and 2.18 ppm. For
deacetylated derivatives 23–28, singlets and triplets re-
ferring to hydroxyl groups were observed, with coupling
constants between 5 and 6 Hz. A typical trans-diaxial
coupling constant (3J 8.0 Hz) was observed for the ano-
meric hydrogen of all final glucosides, which confirms
that all of them were obtained as the β-anomers. In the
13C NMR spectra of derivatives 17–22, signals between
170.6 and 168.0 ppm corresponding to the ester carbonyl
groups were observed. The signals of the saccharide car-
bons were recorded between 85 and 63 ppm for all the
glucosides. The methylene carbon attached to the tri-
azolic ring gave a signal near 70 ppm for triazoles 17–
28 and near 62 ppm for triazoles 29–34. The infrared and
NMR spectra of all synthesized compounds are presented
in the Supplementary Material.
The synthesized 1,2,3-triazoles were also character-
ized by high-resolution mass spectrometry (HRMS),
and the molecular ion peak values are shown in Table
A (Supplementary Material) As could be observed, the
mass accuracy between the calculated and found values
was lower than 5 ppm, which are acceptable for propos-
ing a molecular formula using this technique (Balogh,
2004). The HRMS spectra are also presented in the
Supplementary Material.
Other physicochemical properties of 1,2,3-triazoles
17–34 were determined (solubility and calculated logP)
in order to have a better understanding of the biological
results described below. These data are shown in Table B
(Supplementary Material).
selectivity indexes (SI) were determined by the ratio
CC50 (MRC-5)/IC50 of each strain. The results are shown
in Table 1.
The synthesized compounds did not show significant
cytotoxicity against fibroblast human cells (MCR-5), with
CC50 values varying between 173.8 and 487.6 μM, which
was sufficient to confer interesting SI values for the ac-
tive substances. Considering the antifungal activity, the
peracetylated triazoles 17 (derived from eugenol) and 18
(derived from dihydroeugenol) were the most promis-
ing ones. Triazole 17 was active against three species of
Candida (C. tropicalis, C. krusei, and C. glabrata) with
IC50 values in the range 26.1–52.1 μM. This compound
was four times more potent than fluconazole against
C. krusei, presenting a good selectivity to this species,
although it was not active against C. albicans or C. para-
psilosis. In addition, triazole 17 was less toxic against the
MCR-5 cells (SI > 6.6) than fluconazole (SI > 3.3) con-
sidering C. krusei. The compound 18 showed consider-
able activity against C. tropicalis e C. krusei, as observed
for its analogue 17. This compound was four times more
potent and less toxic than fluconazole against C. krusei
and as active as 17 against C. tropicalis with the same
IC50 value (25.98 μM). The prominent antifungal activ-
ity and spectrum of action found by the derivatives 17
and 18 suggest that the presence of acetyl groups con-
tributed positively to anti-Candida activity, since the
respective deacetylated glycosides (23 and 24) did not
show considerable antifungal action. The peracetylated
triazoles 19, 20, 21, and 22 were inactive against the
evaluated strains. This allows us to conclude that the
presence of an allyl or propyl chain seems to cooperate
significantly to antifungal activity, as seen in the com-
pounds 17 and 18. The conjugation of a double bond
with the aromatic ring also appears to be deleterious for
the antifungal activity of the compounds, since deriva-
tives 21 (formyl group conjugated) and 22 (1-propenyl
group conjugated) were also inactive. Another import-
ant observation about the structure–activity relation-
ship of the synthesized substances is that the presence
of carbohydrate is essential for anti-Candida activity,
since the non-glycosylated derivatives corresponding
to 17 and 18 (29 and 30, respectively) were inactive or
weakly active against some species (IC50 values between
242.7 and 407.9 μM). Finally, when comparing the anti-
Candida activity of glucosyltriazoles 17 and 18 with the
peracetylated glycoside of eugenol already reported by
our group (compound II, Figure 1), we can conclude
that the presence of the 1,2,3-triazole ring contributed
to a large increase in activity against C. tropicalis and
C. krusei, as related triazoles were active at 26.1 μM
against these two strain, while glycoside II was weakly
active at 202 μM.
2.2
Biological assays
|
Following, the compounds were evaluated against dif-
ferent species of fungi (Candida albicans, C. tropicalis,
C. krusei, C. glabrata, and C. parapsilosis) by the micro-
dilution method, and the results were estimated using
the inhibitory concentration that was able to inhibit 50%
of microbial growth (IC50; fungistatic activity; EUCAST
- European Committee on Antibiotic Susceptibility
Testing, 2008). The cytotoxicity of the compounds
was also evaluated against healthy human lung fibro-
blast cells (MRC-5 cells) by MTT colorimetric method
(Mosmann, 1983; Twentyman & Luscombe, 1987). The

MAGALHÃES et al.
5
|
TABLE 1 In vitro antifungal activity (IC50, µM), cytotoxicity against MRC-5 cells (CC50, µM) and selectivity index (SI, CC50/IC50) of the
compounds 17–34
Ca (SI)
IC50
Ct (SI)
IC50
Ck (SI)
IC50
Cg (SI)
IC50
Cp (SI)
IC50
Compound
MRC-5CC50
a
17
–
26.1 (>6.6)
26.1 (>6.6)
52.1 (>3.3)
–
>173.8
18
173.2 (>1)
25.98 (>6.7)
25.98 (>6.7)
103.94 (>1.6)
–
>173.2
19
–
–
–
–
–
>186.8
20
21
–
–
–
–
–
>197.95
>177.5
–
–
–
–
–
22
–
–
173.8 (>1)
–
–
>173.8
23
245.6 (>1)
–
245.6 (>1)
–
245.6 (>1)
>245.6
24
244.3 (>1)
–
–
–
–
>244.3
25
26
–
–
–
–
–
158.7 1.63
>296.6
–
–
–
–
–
27
–
–
–
–
–
>253.1
28
–
–
147.3 (>1.6)
–
–
>245.6
29
30
31
–
–
407.98 (0.52)
–
–
215.6 2.19
265.8 1.33
>487.6
404.6 (0.65)
–
–
404.6 (0.65)
242.7 (1.1)
–
–
–
–
–
32
–
–
–
–
–
–
33
–
–
–
–
–
>429.0
34
–
–
407.9 (0.4)
104.4 (>3.1)
407.9 (0.4)
16 (>20)
–
180.4 1.53
>326.5
Fluconazole
3.2 (>102)
3.2 (>102)
6.4 (>51)
aInactive at highest evaluated concentration; Ca: Candida albicans ATCC 10231; Ct: Candida tropicalis ATCC 750; Ck: Candida krusei ATCC 6258; Cg:
Candida glabrata ATCC 90030; Cp: Candida parapsilosis ATCC 22019; and MRC-5: human lung fibroblasts cells.
FIGURE 3 (a) Comparison
between lowest RMSD solution of VT1
from redocking (cyan sticks) and its
crystallographic conformation (green
sticks). (b) Best docking solutions
of itraconazole (yellow sticks) and
posaconazole (pink sticks). (c) Best
docking solution of peracetylated
compound 21 (blue sticks). (d) Best
docking solution of deacetylated
compound 27 (purple sticks). The heme
prosthetic group is shown in orange
sticks, and some binding site residues are
shown as white sticks with a transparent
solvent accessible surface

6
MAGALHÃES et al.
|
2.3
Molecular modeling studies
docking experiment (VT1 91.8 vs itraconazole 117.9 and
posaconazole 105.7).
|
In order to investigate the likely mechanism of action
of the active derivatives, we decided to perform docking
studies of the synthesized compounds on 14α-lanosterol
demethylase (CYP51). This is a heme-containing enzyme
that is involved in ergosterol synthesis, hence vital for
membrane stability and fungal survival. It is one of the
most explored targets for fungi and has as representative
approved drugs itraconazole, posaconazole, and flucona-
zole (Mani Chandrika & Sharma, 2020). Given the pres-
ence of an azole and an aromatic ring in their structures,
the synthesized compounds in this study bear some struc-
tural similarity with marketed inhibitors of CYP51. It is
known from crystallographic structures that this azole
ring establishes an important interaction with the iron in
the heme cofactor, and this is the reason why it is con-
sidered the pharmacophore of the azole drugs (Hargrove
et al., 2017; Peyton et al., 2015). By performing docking of
the triazole derivatives against CYP51, we sought to inves-
tigate not only the probability of acting on this target, but
also the role of the triazole ring in the predicted complex.
The crystallographic structure used in the molecular
modeling studies (PDB code 5tz1) consisted of a 14α-
lanosterol demethylase from C. albicans, one of the evalu-
ated species, and it was resolved in an adequate resolution
of 2.0 Å. The structure is a complex of the protein with the
inhibitor VT1, also an azole compound, in which, it is pos-
sible to observe a coordination between a nitrogen from
the azole ring and the iron atom from the heme group
(Hargrove et al., 2017).
The docking protocol was determined by choosing
a scoring function, the number of runs, and size of the
search area that reproduced the ligand conformation
closer to the crystallographic complex, given by its root
mean square deviation (RMSD). The value found of 1.03 Å
was lower than the 2.0–2.5 Å usually recommended in the
literature, thus confirming the protocol liability in predict
ligand conformation (Figure 3a). More importantly, the
docking correctly predicted the azole-iron coordination
found in the crystallographic structure.
For referencing purposes, two known azoles that act
in the CYP51 system were also docked: itraconazole and
posaconazole. Both drugs had the same binding mode of
VT1: one nitrogen of the azole ring interacting with the
iron atom in the heme group, a hydrogen bond with a
conserved water and the halide substituted phenyl ring
inserted into a hydrophobic pocket near the heme group
(Figure 3b). Additionally, their protonated piperazine
ring is capable of establishing an additional hydrogen
bond with a conserved water molecule. Altogether, these
observations illustrate the importance of the azole ring
and help to explain the higher Goldscore obtained in the
When analyzing synthesized 1,2,3-triazoles 17–34, the
top3 ranked compounds, with a Goldscore close to VT1’s,
were peracetylated compounds (17, 18, and 21). In the pre-
dicted binding poses, the carbohydrate moiety is placed
near the entrance of the binding site, while the phenyl
ring is pointed toward the heme group. Interestingly, the
carbonylic oxygens of compound 21 found 2.63 Å away
from the heme iron atom, thereby establishing an inter-
action (Figure 3c). Although the deacetylated derivatives
ranked more inferiorly classified than the peracetylated,
hence removal of acetyl groups allowed the molecule to
position itself deeper into the pocket, as observed by a
shift of 3.49 Å when comparing 27 and 21 glucosidic rings
(Figure 3d). Also, some of the free hydroxyls could engage
hydrogen bond with a conserved water in the binding site,
as seen in predicted binding poses of the approved drugs
and the phenyl ring is positioned deeper in the pocket,
establishing hydrophobic contacts. Although these hydro-
gen bonds contributed to their final scoring, peracetylated
larger van der Waals contribution and water displace-
ment were more relevant, as the binding pocket is formed
mostly by hydrophobic residues. However, deacetylated
derivatives might play an important role in antifungal
activity, since the peracetylated compounds could be me-
tabolized once absorbed by fungi to desacetylated corre-
sponding derivatives.
Despite their lower Goldscore (<63.69), monosubsti-
tuted triazoles 29 and 30 had a similar interaction mode
than azole drugs regarding the heme group. The low score
and antifungal activity shown by these non-glycosylated
derivatives could be associated with their lower molecular
weight, as less atoms are available to interact with the tar-
get; therefore, further optimization should be considered
to try to improve their action. For instance, compound 30
had a discreet activity against C. parapsilosis and it has a
reasonable logP for improved cellular penetration (calcu-
lated logP 2.62).
Since docking has some intrinsic limitations, such as
treating the protein as a rigid body and the absence of an
explicit water model, molecular dynamics (MD) were per-
formed in order to refine docking results and observe the
binding dynamics. Since this method has a higher compu-
tational cost when compared to docking, we choose rep-
resentative molecules for this study: the crystallographic
ligand (VT1), the best ranked peracetylated (21) and
deacetylated (27), the best-monosubstituted triazole (30),
the most active compound (18) and its deacetylated de-
rivative (24). We also decided to include an inactive com-
pound in the studies. For this purpose, we searched the
ChEMBL database for inactive compounds tested against
C. albicans 14α-lanosterol demethylase and selected

MAGALHÃES et al.
7
|
compound 13758 for being more structurally related to
synthesized compounds (Figure 4).
during the trajectory indicating that docking failed to cor-
rectly predict these complex conformations, resulting in
non-stable ones (Figure 5f).
The complex stability was confirmed during the equil-
ibration phase by monitoring the fluctuation of the tem-
perature, pressure, and density of the system. Also, in
all complexes, the protein radius of gyration (Rg), an in-
dication of protein compactness, was maintained stable
during the entire 20 ns production simulation and a simi-
lar RMSF profile was observed for all complexes (Figure 5;
Lemkul, 2019).
The protein in all the complexes had only small con-
formation changes throughout the simulation, given by
their mean heavy atoms RMSD (1.5–1.9 Å) to the crystal-
lographic structure. These changes occurred mostly at the
beginning of the simulation, since the complex is going
from a crystal packing to a solvated system.
As for the ligands, when compared to the initial con-
formation from docking, VT1 and 30 had the least varia-
tion, with a mean RMSD below 2.0 Å, while compound 24
had an acceptable variation (2.5 Å). In the contrary, com-
pounds 18, 21, and 27, as well as the inactive compound
13758, had significative variations in their conformations
Although ligands with hydrophobic character are wel-
come in this particular binding site, it seems reasonable
to think that extra volume conferred by acetyl groups
of peracetylated compounds might be the cause of such
conformation changes. When comparing the trajecto-
ries of peracetylated compound 18 and its correspondent
deacetylated 24, there is a clear difference in the binding
dynamics. While 18 moves outwards of the hydrophobic
channel, thus leaving the binding pocket, compound 24
has a lateral movement, placing its phenyl ring deeper in
the hydrophobic pocket near heme-binding site (Figure 6).
Although peracetylated compounds 17 and 18 per-
formed better than correspondent deacetylated in the in
vitro assays, the possibility of these compounds act as pro-
drugs might help to explain such divergence. It is known
that ester groups could be readily hydrolyzed once in-
side fungal cells, thus affording deacetylated compounds
which in turn can bind to the molecular target in the cy-
toplasmic side of plasma membrane (Monk et al., 2014).
Consequently, compounds 17 (logP: 3.34) and 18 (logP:
3.25), having a presumably better absorption profile given
by their calculated logP, would penetrate the fungal to af-
ford their correspondent deacetylated active compounds.
On the other side, when tested directly, deacetylated 23
(logP: 0.53) and 24 (logP: 0.61) are too much hydrophilic
to cross the fungal membrane, being inactive in the in vitro
assays.
As coordination between azole ring and the heme iron
atom is an important feature in ligand binding, we also
FIGURE 4 Structure of inactive compound 13758
FIGURE 5 10ps moving averages of parameters monitored during the equilibration phase: temperature (a), pressure (b), and density
(c); and the protein radius of gyration in the MD simulation (d). VT1 (black), 17 (red), 21 (blue), 24 (green), 27 (yellow), 30 (purple), and
13758 (gray)

8
MAGALHÃES et al.
|
(a)
(b)
FIGURE 6 Trajectory snapshots of 18 (a) and 24 (b). The snapshots were extracted at each 1 ns of simulation and the ligand gradient
colored from red (first frame) to blue (last frame). A referencing protein is partially shown as a yellow ribbon and the heme group as orange
sticks with the iron atom highlighted as a green sphere. The arrow indicates the tendency of the ligand movement in the binding site
FIGURE 8 Solvent-accessible surface area of binding site
residues throughout the simulation
FIGURE 7 Interatomic distance between triazole nitrogen
(VT1 and 30) or carbonylic oxygen (21 and 27) and the iron atom
throughout the simulation. VT1 (black), 21 (blue), 27 (yellow), and
30 (purple)
water molecules in the binding site (Alonso et al., 2006).
For reference, the residues of the binding site in the VT1
complex had a mean SASA of 2.33 nm2, while 18 and 21
had a SASA of 2.11 and 2.73 nm2, respectively (Figure 8).
The complex with compound 24 had a SASA of 2.26 nm2,
close to the crystallographic ligand, despite its lower mo-
lecular weight (Figure 8). In CYP51, a binding site which
is formed mostly by hydrophobic residues, where the dis-
placement of highly organized water could favor complex
formation by a gain in the system entropy (Alonso et al.,
2006). On the other side, the complexes with compounds
27 and 30 had a calculated SASA closer to complex with
the inactive compound 13758 (3.81, 3.27, and 4.00 nm2, re-
spectively), meaning that the binding site is not effectively
occupied by these ligands (Figure 8).
The total interaction energy of the complex calcu-
lated by Gromacs (sum of Coulombic and Lennard-Jones
terms) could provide an estimate of complex affinity, al-
though other parameters should be considered to deter-
mine the ligand binding energy, such as the solvation
effects. Considering the last 15 ns of simulation, most
compounds had a total interaction energy close to VT1
measured the distance between ligand triazole nitrogen
and the iron throughout the simulation (Hargrove et al.,
2017; Lepesheva et al., 2018; Lepesheva & Waterman,
2004). Both VT1 and 30 maintained a mean distance be-
tween these atoms compatible with a coordination bond
(VT1 N-Fe: 2.6 Å and 30 N-Fe: 2.5 Å; Figure 7; Kirton
et al., 2005). As for the other ligands, the phenyl ring was
turned toward the heme group while the triazole ring was
too far away to establish any interaction. Though 21 and
27 could engage such interaction via their carbonylic oxy-
gen, only in the complex with 27 this interaction was ob-
served, although only in the first 5 ns, when a significant
conformational change occurred (Figure 7).
The solvent-accessible surface area (SASA) is an im-
portant measure when analyzing ligand binding and com-
plex stability. A relatively lower SASA means a greater
complementarity between ligand and the biding site and
a large difference between SASA in the bounded and
nonbonded complex could indicate the displacement of

MAGALHÃES et al.
9
|
FIGURE 9 Representative conformations of 24 (green sticks) and 27 (yellow sticks) extracted from final 1ns of MD after clustering. The
heme prosthetic group is shown in orange sticks; water molecule oxygens are shown as red-scaled balls and important binding site residues
as white sticks. Only the main non-bonding interactions between CYP51-24 are shown as dashed lines (green, hydrogen bond; pink, π-
stacking) and distances in angstroms for clarity. A solvent-accessible surface was of residues Thr122, Phe126, Ile131, Leu300, and Ile304 is
shown to emphasize the hydrophobic pocket
(−299.06 kcal/mol), while the inactive 13758 and 30 had
the worst interaction energy (−184.09 and −173.28 kcal/
mol, respectively). Surprisingly, compound 24 had an IE
of −390.96 kcal/mol, value 30% lower than VT1, indicat-
ing a greater affinity.
methoxy group acts as a hydrogen bond donor in the in-
teraction with Tyr132. A similar interaction is observed
with VT1 and in other crystallographic complexes (i.e.,
5EAD and 5ESM), but with the alcoholic hydroxyl and
mediated by water.
Visualization of both trajectories reveals that propyl
(24) and carbonyl (27) groups end up in a hydropho-
bic pocket formed by residues Thr122, Phe126, Ile131,
Leu300, and Ile304 (Figure 9). As this pocket surrounds
heme carboxylates, repulsion with the carbonylic oxygen
of 27 might be the cause of positive coulombic compo-
nent in the complex. When analyzing each term of the
interaction energy separately, the major contribution
interaction energy of 24 with aforementioned residues
comes from the Lennard-Jones, while the Coulombic
term is close to zero, meaning a strong hydrophobic
contribution to the interaction energy. In relation to 27,
the Coulombic contribution is larger at first and turns
repulsive at the end of simulation, due to proximity to
heme carboxylates. Also, the calculated SASA of these
residues for complex with 24 has the same value of VT1
(0.12 nm2), while in the complex with 27 the SASA is
more than doubled (0.28 nm2).
In Figure 9 is also possible to visualize the main
nonbonded interactions between 24 and the protein. In
addition to those arising from insertion of propyl sub-
stituent in the hydrophobic pocket, a π-stacking inter-
action is observed between triazole ring and the Tyr118
residue. Several hydrogen bonds were also observed,
mainly from saccharidic portion, with some of them in-
volving water molecules. Interesting, the oxygen in the
3
CONCLUSION
|
In this research study, eighteen new glucosyl-1,2,3-
triazoles derived from eugenol and analogues were
synthesized and characterized by spectroscopy and spec-
trometry methods. The new compounds were evaluated
against Candida spp. by microdilution method, and the
promising ones were peracetylated derivatives 17 (an
eugenol derivative) and 18 (a dihydroeugenol deriva-
tive). Furthermore, the triazole 17 was active C. tropi-
calis, C. krusei, and C. glabrata at 26.1–52.1 μM, and it
was four times more potent than fluconazole against
C. krusei. Derivative 18 showed a wider spectrum of ac-
tivity than 17, and it was four times more potent and less
toxic than fluconazole against C. krusei and as active as
17 against C. tropicalis. Altogether, molecular modeling
studies suggest that antifungal activity of compounds
17 and 18 might be due to inhibition of CYP51, thereby
disrupting fungal ergosterol synthesis. Since their re-
spective deacetylated derivatives, 23 and 24, showed to
be more promising but had lower activity in vitro, we
believe that peracetylated compounds might be acting
as prodrugs. Guided by the molecular modeling results,
further modifications in the glycosidic ring and the R1

10
MAGALHÃES et al.
|
via a cascade including propargyl azides, allenyl azides, and
triazafulvenes. Arkivoc, 1, 338–361. https://doi.org/10.24820/
ark.5550190.p009.846
substituent at the phenyl ring should improve ligand
binding. Additionally, despite their lower activity, mono-
substituted triazoles showed a similar binding mode of
known active azoles and could be used as a starting point
in the design of new ligands.
Bhattacharya, S., Sae-Tia, S., & Fries, B. C. (2020). Candidiasis and
mechanisms of antifungal resistance. Antibiotics, 9, 1–19.
https://doi.org/10.3390/antibiotics9060312
EUCAST - European Committee on Antibiotic Susceptibility Testing.
(2008). Method for the determination of minimum inhibitory
concentration (mic) by broth dilution of fermentative yeasts.
Definitive document E.Def. 7.2, European Society of Clinical
Microbiology and Infectious Diseases, 2008.
Guilherme, F. D., Simonetti, J. E., Foolquitto, L. R. S., Reis, A. C. C.,
Oliver, J. C., Dias, A. L. T., Dias, D. F., Carvalho, D. T., Brandão,
G. C., & Souza, T. B. (2019). Synthesis, chemical characterization
and antimicrobial activity of new acylhydrazones derived from
carbohydrates. Journal of Molecular Structure, 1184, 349–356.
Hargrove, T. Y., Friggeri, L., Wawrzak, Z., Qi, A., Hoekstra, W. J.,
Schotzinger, R. J., York, J. D., Peter Guengerich, F., & Lepesheva,
G. I. (2017). Structural analyses of Candida albicans sterol 14α-
demethylase complexed with azole drugs address the molecu-
lar basis of azole-mediated inhibition of fungal sterol biosyn-
thesis. Journal of Biological Chemistry, 292, 6728–6743. https://
doi.org/10.1074/jbc.M117.778308
Kirton, S. B., Murray, C. W., Verdonk, M. L., & Taylor, R. D. (2005).
Prediction of binding modes for ligands in the cytochromes
P450 and other heme-containing proteins. Proteins: Structure,
Function, and Bioinformatics, 58, 836–844. https://doi.
org/10.1002/prot.20389
ACKNOWLEDGMENTS
The authors acknowledge the PROPP/UFOP, FAPEMIG
(APQ-00820-19), and CNPq agencies for the scholarships
and financial support provided. The authors acknowl-
edge the “Laboratório Multiusuário de Caracterização de
Moléculas (Pharmacy School at the Federal University
of Ouro Preto)" for nuclear magnetic resonance analy-
sis. The authors acknowledge the “Base de Estruturas
Cristalinas/CAPES (Crystalline Structures Base)” for
the GOLD software licensing. The authors would also
like to acknowledge the “Laboratório Multiusuário de
Proteômica e Biomoléculas (LMU-ProtBio), do Núcleo
de Pesquisas em Ciências Biológicas” (from the Research
Nucleus in Biological Sciences), Federal University of
Ouro Preto, MG, Brazil, for providing the required equip-
ments and technical expertise for the sample processing
and analyses.
CONFLICT OF INTEREST
The authors declare no conflict of interest.
Lemkul, J. (2019). From proteins to perturbed hamiltonians: A suite
of tutorials for the GROMACS-2018 molecular simulation pack-
age [Article v1.0]. Living Journal of Computational Molecular
Science, 1, 1–53. https://doi.org/10.33011/livecoms.1.1.5068
Lepesheva, G. I., Friggeri, L., & Waterman, M. R. (2018). CYP51 as
drug targets for fungi and protozoan parasites: Past, present and
future. Parasitology, 145, 1820–1836. https://doi.org/10.1017/
S0031182018000562
DATA AVAILABILITY STATEMENT
Data sharing is not applicable to this article as all data have
been made available in the “Supplementary Material”
section.
Lepesheva, G. I., & Waterman, M. R. (2004). CYP51 - The omnipo-
tent P450. Molecular and Cellular Endocrinology, 215, 165–170.
https://doi.org/10.1016/j.mce.2003.11.016
Mani Chandrika, K. V. S., & Sharma, S. (2020). Promising antifungal
agents: A minireview. Bioorganic & Medicinal Chemistry, 28,
115398. https://doi.org/10.1016/j.bmc.2020.115398
Monk, B. C., Tomasiak, T. M., Keniya, M. V., Huschmann, F. U.,
Tyndall, J. D. A., O’Connell, J. D., Cannon, R. D., McDonald,
J. G., Rodriguez, A., Finer-Moore, J. S., & Stroud, R. M. (2014).
Architecture of a single membrane spanning cytochrome P450
suggests constraints that orient the catalytic domain relative to
a bilayer. Proceedings of the National Academy of Sciences, 111,
3865–3870. https://doi.org/10.1073/pnas.1324245111
Mosmann, T. (1983). Rapid colorimetric assay for cellular growth
and survival: Application to proliferation and cytotoxicity as-
says. Journal of Immunological Methods, 65, 55–63.
ORCID
Diogo Teixeira Carvalho https://orcid.
org/0000-0002-6362-2822
Danielle Ferreira Dias https://orcid.
org/0000-0001-9129-4734
Thiago Belarmino de Souza https://orcid.
org/0000-0003-3627-6407
REFERENCES
Achard, T., Lepronier, A., Gimbert, Y., Clavier, H., Giordano, L.,
Tenaglia, A., & Buono, G. (2011). A regio- and diastereoselec-
tive platinum-catalyzed tandem [2+1]/[3+2] cycloaddition
sequence. Angewandte Chemie, 123, 3614–3618. https://doi.
org/10.1002/ange.201007992
Alonso, H., Bliznyuk, A. A., & Gready, J. E. (2006). Combining
docking and molecular dynamic simulations in drug de-
sign. Medicinal Research Reviews, 26, 531–568. https://doi.
org/10.1002/med.20067
Balogh, M. P. (2004). Debating resolution and mass accuracy in mass
spectrometry. Spectroscopy, 19, 34–40.
Peyton, L. R., Gallagher, S., & Hashemzadeh, M. (2015). Triazole
antifungals: A review. Drugs of Today, 51, 705–718. https://doi.
org/10.1358/dot.2015.51.12.2421058
Pote, A. R., Vannam, R., Richard, A., Gascón, J., & Peczuh, M. W. (2018).
Formation of and glycosylation with per-O-acetyl septanosyl ha-
lides: rationalizing somplex reactivity en route to p-nitrophenyl
septanosides. European Journal of Medicinal Chemistry, 14,
1709–1719. https://doi.org/10.1002/ejoc.201800310
Banert, K., Hagedorn, M., Hemeltjen, C., Ihle, A., Weigand, K., &
Priebe, H. (2016). Synthesis of N-unsubstituted 1,2,3-triazoles

MAGALHÃES et al.
11
|
Revie, N. M., Iyer, K. R., Robbins, N., & Cowen, L. E. (2018).
Antifungal drug resistance: Evolution, mechanisms and im-
pact. Current Opinion in Microbiology, 45, 70–76. https://doi.
org/10.1016/j.mib.2018.02.005
Safavi, M., Asthari, A., Khalili, F., Mirfazli, S. S., Saeedi, M.,
Ardestani, S. K., Ranjbar, P. R., Tehrani, M. B., Larijani, B., &
Mahdavi, M. (2019). Novel quinazolin-4(3H)-one linked to
1,2,3-triazoles: Synthesis and anticancer activity. Chemical
Biology & Drug Design, 92, 1373–1381. https://doi.org/10.1111/
cbdd.13203
Souza, T. B., Brito, K. M. O., Silva, N. C., Rocha, R. P., Sousa, G.
F., Duarte, L. P., Coelho, L. F. L., Dias, A. L. T., Veloso, M. P.,
Carvalho, D. T., & Dias, D. F. (2016). New eugenol glucoside-
based rerivative shows fungistatic and fungicidal activity
against opportunistic Candida glabrata. Chemical Biology &
Drug Design, 87, 83–90.
Souza, T. B., Orlandi, M., Coelho, L. F. L., Malaquias, L. C. C., Dias,
A. L. T., Carvalho, R. R., Silva, N. C., & Carvalho, D. T. (2014).
Synthesis and in vitro evaluation of antifungal and cytotoxic ac-
tivities of eugenol glycosides. Medicinal Chemistry Research, 23,
496–502. https://doi.org/10.1007/s00044-013-0669-2.
and chemosensitivity. British Journal of Cancer, 56, 279–285.
https://doi.org/10.1038/bjc.1987.190
Whaley, S. G., Berkow, E. L., Rybak, J. M., Nishimoto, A. T., Barker,
K. S., & Rogers, P. D. (2017). Azole antifungal resistance in
Candida albicans and emerging non-albicans Candida spe-
cies. Frontiers in Microbiology, 7, 1–7. https://doi.org/10.3389/
fmicb.2016.02173
SUPPORTING INFORMATION
Additional supporting information may be found online
in the Supporting Information section.
How to cite this article: Magalhães, L. S. D., Reis,
A. C. C., Nakao, I. A., Péret, V. A. C., Reis, R. C. F.
M., Silva, N. C., Dias, A. L. T., Carvalho, D. T., Dias,
D. F., Brandão, G. C., Braga, S. F. P., & Souza, T. B.
D. (2021). Glucosyl-1,2,3-triazoles derived from
eugenol and analogues: Synthesis, anti-Candida
activity, and molecular modeling studies in CYP-51.
Chemical Biology & Drug Design, 00, 1–11. https://
doi.org/10.1111/cbdd.13948
Strickland, A. B., & Shi, M. (2021). Mechanisms of fungal dissem-
ination. Cellular and Molecular Life Sciences, 78, 3219–3238.
https://doi.org/10.1007/s00018-020-03736-z
Twentyman, P. R., & Luscombe, M. (1987). A study of some vari-
ables in a tetrazolium dye (MTT) based assay for cell growth