Molecular docking evaluation of (E)-5-arylidene-2-thioxothiazolidin-4-one derivatives as selective bacterial adenylate kinase inhibitors

Article abstract of DOI:10.3390/molecules23051076

Multi-drug resistant microorganism infections with emerging problems that require not only a prevention strategy, but also the development of new inhibitory compounds. Six previously synthesized 5-arylidene-2-thioxothiazolidin-4-one derivatives 1a–f, were

Full text of DOI:10.3390/molecules23051076

molecules
Article
Molecular Docking Evaluation of
(E)-5-arylidene-2-thioxothiazolidin-4-one Derivatives
as Selective Bacterial Adenylate Kinase Inhibitors
Mihaela Ileana Ionescu 1,2,* and Ovidiu Oniga 3
1
Department of Microbiology, Iuliu Hatieganu University of Medicine and Pharmacy,
,
400349 Cluj-Napoca, Romania
Department of Microbiology, County Emergency Clinical Hospital, 400006 Cluj-Napoca, Romania
Department of Pharmaceutical Chemistry, Faculty of Pharmacy, Iuliu Ha,
2
3
t
ieganu University of Medicine and
Pharmacy, 400012 Cluj-Napoca, Romania; onigao65@yahoo.com
Correspondence: mionescu@umfcluj.ro
*
Academic Editor: Thierry Besson
ꢀꢁꢂꢀꢃꢄꢅꢆꢇ
ꢀꢁꢂꢃꢄꢅꢆ
Received: 21 March 2018; Accepted: 1 May 2018; Published: 3 May 2018
Abstract: Multi-drug resistant microorganism infections with emerging problems that require not
only a prevention strategy, but also the development of new inhibitory compounds. Six previously
synthesized 5-arylidene-2-thioxothiazolidin-4-one derivatives 1a–f, were screened for inhibitory
activity on adenylate kinases of different origins by molecular docking. The compounds 1c and 1d
were the most efficient inhibitors of bacterial and some archean adenylate kinases. Hydrogen bond
interactions were observed with the residues belonging to the ATP binding site. Moreover human
adenylate kinases are poor targets, suggesting that this selectivity offers promising prospectives for
refining the structure of our compounds.
Keywords: adenylate kinase; thiazolidine; molecular docking; inhibitory activity
1. Introduction
Sulfur is a non-metal widespread in Nature both as the pure element or as sulfide or sulfate
minerals, while sulfur dioxide (SO2) and hydrogen sulfide (H2S) are abundantly found in the gases
emitted by volcanoes. The ancient people, good observers of the world, not only discovered the
healing properties of sulfurous springs, but in Roman mythology, along myriads of deities, there was
a goddess Mephitis responsible for the unpleasant smell of these environments. Sulfur-containing
compounds are widely found in Nature and climatic changes, which have occurred during billions of
years, greatly influence the chemical composition of soil. Consequently, all living things must have
evolved in parallel with inorganic and organic compounds found in each specific ecologic niche [1–3].
The sulfur is essential for human health and sulfur-containing compounds could influence human
health in both senses—beneficially or related with the most diverse pathologies [4,5].
The pharmaceutical industry has developed a huge panel of sulfur-containing drugs with
anti-rheumatic, oncological or anti-inflammatory effects [
6,7]. New compounds with antibacterial
effect are continuously being developed [ ]. Nowadays there is a great interest in finding natural
8,9
compounds with therapeutic effect, such garlic extract [10]. Although it seems there is a low
interest towards synthesis of new sulfur-containing heterocycles—only 382 results, on 17 March 2018,
when searching “sulfur-containing AND heterocycles” on Clarivate Web of Science—the biological
implications of some of them could reveal unexpected applications. Focusing on a “thiazolidine” search,
among 2661 results, after refining search with the query “clinical trials” only 22 papers were selected.
While the health care professionals are always willing to start clinical trials, the new sulfur-containing
Molecules 2018, 23, 1076; doi:10.3390/molecules23051076
www.mdpi.com/journal/molecules

Molecules 2018, 23, 1076
2 of 11
heterocycles remains in the synthesis and characterization phase. Finding new compounds implies
accurate target identification. Thiazolidine derivatives show antibacterial or antifungal activity,
inhibiting the most diverse structural or functional enzymes. A recent experimental study showed
that some 1,2,4-triazolothiazolidin-4-one derivatives perturb the cellular integrity by inhibition of
an enzyme involved in peptidoglycan synthesis— UDP-N-acetylenolpyruvoylglucosamine reductase
(MurB) [11]. Other authors identified Mycobacterium tuberculosis pantothenate synthetase as a target for
thiazolidine derivatives [12]. Thiazolidine derivatives definitely, have interesting features that are not
sufficiently explored, but the real concern is related to their side effects on human enzymes.
Enzymes involved in energetic metabolism, as nucleoside monophosphate kinases, have two
main characteristics—ubiquity and cruciality for life [13–15]. Adenylate kinase (AK)—ATP:
AMP phosphotransferase; EC 2.7.3.4—catalyzes the reversible transfer of the phosphate group between
ATP and AMP (Mg2+ ATP + AMP = Mg2+ ADP + ADP). Some isoenzymes—AK3 and AK4, also named
AK3-like—are GTP:AMP phosphotransferase; EC 2.7.4.10 [16–18]. It is a usually monomeric enzyme
that has been chosen as a model in many kinetic studies, its conformational changing during catalytic
reactions being revealed by crystallographic studies of the unliganded enzyme or of the enzyme in
complex with the substrate analog Ap5A [19]. It has a three-domain structure which undergoes closure
through a hinge mechanism. The large central domain, named CORE domain, is flanked by two
small domains—the AMP binding domain (AMPbd) and the ATP-binding domain (ATPlid) or LID
domain [20]. The latter closes over the site of phosphoryl transfer upon ATP binding [21], resulting
in a so-called closed conformation [22,23]. The conformational rearrangement process of the ATPlid
domain and, to a smaller measure, of the AMPbd domain is not clearly understood. Experimental
studies show that the flexible domains fold in a noncooperative manner with respect of the CORE
domain [24]. This conformational pathway is not a steep one. Between the open and the closed
conformations, there are also intermediate states. Coarse-grained molecular dynamics simulation
studies provide insight into the residues involved in structural fluctuation in the open, closed, and
intermediate forms [25]. The large-conformational movement of the ATPlid, at the beginning of
catalytic process, has a much more complex role than just proper alignment of the substrates in the
active sites. It is demonstrated that the catalytic selectivity for ATP over GTP lies on the ability of the
same residues—Arg 118, Arg 123, and Arg 167—to influence the positive selection of ATP and the
negative selection of GTP. Therefore, energy homeostasis is maintained by reducing the depletion of the
cellular GTP pool [26]. AK, a ubiquitous enzyme, shows remarkable adaptation over billions of years
of evolution from a hot environment to a cooler temperature, the central CORE domain determining
the melting temperature of the AKs. [24]. Therefore it is not a surprise that it is present in all three
domains of life—Bacteria, Eukarya, and Archaea. Until now, nine AK isoforms were identified.in
eukaryotes distributed in different intracellular compartments. While AKs isoenzymes 1, 5, 7 and
8 are cytosolic, AK6 and AK 9 are found in the nucleus. AK2, AK3, and AK4 are mitochondrial
isoenzymes—AK2 being located in the intermembrane space [27,28]. Some studies demonstrated the
role of AK of Burkholderia pyrrocinia in expressing its antifungal activity [29]. It is well established that
elemental sulfur is a specific inhibitor of AN activity from skeletal muscle, erythrocytes, and cardiac
tissue [30] and sulfhydryl compounds promote conformational changes of AKs, a feature which could
be exploited in controlling the inhibitory activity of other compounds [31].
All these facts presented above led us to consider AK as an interesting choice for testing new
sulfur-containing heterocycles. In a previous work, a series of 5-arylidene-2-thioxothiazolidin-4-one
derivatives demonstrated inhibitory activity on AKs from Escherichia coli and Streptococcus pneumonia.
All derivatives were initially tested at one concentration (0.2 mM), then the best inhibitor—herein
named 1d—was selected for further kinetic analysis. The I50 values (the inhibitor concentration
that leads to 50% activity inhibition) were 0.067 mM for E. coli AK and 0.065 mM for S. pneumoniae,
respectively [32,33]. We assumed that the binding site of this derivative is the same for AKs of both
origin. To decipher the binding of this series of 5-arylidene-2-thioxothiazolidin-4-one derivatives,
a different approach was attempted. Even though considered an empirical method, molecular docking

Molecules 2018, 23, 1076
3 of 11
is a useful tool to gain insights into ligand-enzyme interactions and for screening AKs of different
origin, mostly comparing AKs of bacterial pathogens with their human counterparts.
2. Results
First, the raw data—binding energy ∆G (kcal/mol) and inhibition constant Ki (µM) of the best
conformation after molecular docking—were analyzed by statistical tests. We attempt to organize the
data started by the hypothesis that there is no difference between inhibitory activities of the derivatives
1a–f for AKs of bacterial or human origin, but, before that, we have to find out if derivatives 1a–f
interact with the active site residues of AKs. Therefore, the 3D structures of unliganded AKs—open
conformation—were compared with their closed-conformation counterparts. In Table 1 we selected
the binding energy and the inhibition constant values only for the AKs that have 3D structured in
the Protein Data Bank (PDB) in both conformations. In our previous experimental studies [32,33],
an extra compound was tested—C10H6FNOS2, with the substituent R = 3F—but we failed to docked it.
We were also unable to perform the docking of 4NTZ with 1b and 1c, 1Z83 with 1d and 1f, and 2AR7
with 1b, 1d and 1e, respectively.
Table 1. The binding energies (kcal/mol) and inhibition constants (
µM) of the unliganded AKs or of
the enzyme in complex with the substrates or substrate analogs.
∆G 2 (kcal/mol)
Ki 4 (µM)
Mean (±SD)
p-Value
AK Conformation
(PDB ID 1
Strain
Mean (±SD 3
)
)
∆G (kcal/mol)
Ki (µM)
O 5a (4AKE)
C 5b (1HPQ)
−5.26 (±1.7)
51.32 (±27.53)
E. coli
0.020285
0.002378
−7.19 (±0.24)
5.75 (±2.59)
O (4NTZ)
C (4NU0)
−5.59 (±2.48)
−6.97 (±0.42)
18.12 (±14.01)
9.23 (±4.55)
S. pneumoniae
0.2097
0.177056
0.000854
O (2RH5)
C (2RGX)
−5.81 (±0.32)
59.53 (±22.31)
A. aeolicus
0.000898
−6.77 (±0.4)
13.21 (±9.37)
O (1NKS/chain A)
O/C (1NKS/Chain C)
C (1NKS/Chains F)
−5.87 (±0.09)
−5.91 (±.44)
−6.46 (±0.31)
50.61 (±7.75)
56.04 (±32.03)
34.77 (±14.19)
S. acidocaldarius 6
0.000975
0.000341
O (1KHT/Chain A)
C (1KHT/Chain B)
−5.69 (±0.35)
81.06 (±65.92)
M. voltae
0.000054
0.056913
0.010154
0.024538
0.034455
0.020725
−6.88 (±0.26)
9.73 (±4.31)
O (2AR7)
C (BBW)
−6.11 (±0.22)
−6.52 (±0.27)
34.48 (±13.15)
17.85 (±6.62)
Human–AK4
O (2BWJ/Chain A)
C (2BWJ/Chain B)
−6.41 (±0.27)
22.19 (±13.11)
Human–AK5
1
−5.77 (±0.41)
70.03 (±40.66)
Protein Data Bank Accession number; 2 Binding Energy; 3 Standard Deviation; 4 Inhibition Constant; 5a O = Open
5b
6
conformation of AK,
F comparison.
C = Closed-conformation of AK; p-value for 1NKS/chain A and 1NKS/Chains
In Table S1 (Supplementary Material) we compare the overall inhibitory activity of the derivatives
1a–f against Bacteria, Eukarya (human), and Archaea AKs. Since only two archaean AKs, were included
in the present study, data obtained after molecular docking was not statistically analyzed by Student’s
t-test. However, we performed a statistical analysis of the binding energy and the inhibition constant
values of all AKs in closed conformation to have a global picture of the results. Statistical analysis of
the data reveals that there is enough evidence to indicate that the means are not the same. Because
the One-Way ANOVA test relies on the variability between the populations, a second statistical
test was performed. The Kruskal–Wallis H test confirmed the results obtained by the previous test.
Then, after a detailed analysis of the raw data and the statistical interpretation of molecular docking
parameters, we attempted to understand the intrinsic mechanisms of AK inhibition by derivatives
1a–f. From the Bacteria domain we selected the AKs from the mesophilic Escherichia coli (3HPQ),
Streptococcus pneumoniae (4NU0) and Bacillus subtilis (1P3J), the thermophilic Bacillus stearothermophilus
(1ZIP), the hypertermophilic Aquifex aeolicus (2RGX), and the psychrophilic Bacillus globisporus
(1S3G). Five isoforms of human origin were found as having the 3D structure co-crystallized with
inhibitors or similar substrates—the AK1 (1Z83 and 2C95), AK2 (2C9Y), AK4 (2BBW) and AK5

Molecules 2018, 23, 1076
4 of 11
(2BWJ-chain B). Only two AKs from archean species were selected—the mesophilic Methanococcus
voltae (1KHT/chain B) and hypertermophilic Sulfolobus acidocaldarius (1NKS/chain F). Table 2 shows
the statistical analysis of inhibitory activity of the derivatives 1a–f against bacterial and human AKs in
closed-conformations—the G (kcal/mol) and Ki ( M) data are shown in the Table S1 as mean and
∆
µ
standard deviation. We noted that the most significant differences between inhibitory activity against
AKs of bacterial and human origin are for the compound 1c, therefore we have focused on observing
the residues involved in binding.this compound. Since these observations are not consistent with
our previous experimental studies, when the 1d derivative demonstrated the best inhibitory activity
against E. coli and S. pneumoniae, a thorough analysis of the raw docking data was performed.
Table 2. Statistical analysis of the inhibitory activity of the derivatives 1a–f against closed-conformation
of bacterial AKs (2RGX, 3HPQ, 4NU0, 1P3J, 1ZIP, 1S3G) and eukaryotic AKs (1Z83, 2C95, 2C9Y, 2BBW,
2BWJ/chain B).
1
2
3
Derivative
1a
∆G/Ki
t-Value
p-Value
∆µ
CI 95%
∆G
Ki
0.09867
1.06499
0.923565
0.314617
0.0357
18.4687
−0.7821 ≤ ∆µ ≥ 0.8534
−20.7608 ≤ ∆µ ≥ 57.6981
∆G
Ki
3.84529
2.4127
0.003935
0.039076
0.933
46.6247
0.3841 ≤ ∆µ ≥ 1.4819
2.9091 ≤ ∆µ ≥ 90.3402
1b
1c
1d
1e
1f
∆G
Ki
4.49226
2.2898
0.001506
0.04779
0.958
33.761
0.4756 ≤ ∆µ ≥ 1.4404
0.4075 ≤ ∆µ ≥ 67.1145
∆G
Ki
1.8825
1.87666
0.09634
0.097405
0.66
21.2967
−0.1485 ≤ ∆µ ≥ 1.4685
−4.8723 ≤ ∆µ ≥ 47.4
∆G
Ki
2.31486
1.53788
0.04587
0.158459
0.8183
58.1810
0.0186 ≤ ∆µ ≥ 1.6180
−27.4010 ≤ ∆µ ≥ 143.7630
∆G
Ki
1.08505
1.21529
0.309513
0.258901
0.325
27.4042
−0.3657 ≤ ∆µ ≥1.0157
−24.5951 ≤ ∆µ ≥ 79.4034
1
∆
G (kcal/mol), Ki (µ
M); 2 the difference between eukaryotic mean and bacterial mean; 3 95% confidence interval
of the difference between the eukaryotic mean and bacterial mean.
In Table 3 and Figure 1 a comparison of the inhibitory activity of the derivatives 1c and 1d
demonstrate that their interactions are nor the same, although the experimental data suggest a similar
inhibitory mechanism. However, for E. coli AK, 1c and 1d interactions are comparable, as shown in the
Tables S2 and S3.
Table 3. Analysis of the inhibitory activity of the derivatives 1c and 1d against open and
closed-conformations of E. coli and S. pneumoniae AKs.
1c
1d
Strain
PDB ID
∆G 1
Ki 2
∆G
Ki
4AKE
3HPQ
−5.82
54.17
4.97
−6.65
13.26
4.68
E. coli
−7.24
−7.27
4NTZ
4NT0
−1.91
−7.13
37.12
5.92
−7.33
−7.74
4.22
2.13
S. pneumoniae
∆G (kcal/mol); 2 Ki (µM).
1

Molecules 2018, 23, 1076
5 of 11
3HPQ
E. coli
4NU0
S. pneumoniae
Figure 1. The interactions of the 1c and 1d derivatives with E. coli and S. pneumoniae AKs.
A deeper insight of ligand-protein interactions for the compounds 1c and 1d reveals more evidence
of its efficiency in distinct inhibition of bacterial and some archean AKs. In Tables S2 and S3, Figures
S1 and S2 we show the atomic interactions of the best conformations obtained by molecular docking of
compounds 1c and 1d with different AKs. The AKs’ interacting residues with AMP, ATP, or ADP were
provided the Protein Features module of the PDB. Hydrogen bond interactions were observed with
the residues from the ATPlid for bacterial AKs with a subtle exception for S. pneumoniae AK which
forms a hydrogen bond with a residue belonging to AMPbd. The hypertermophilic bacteria A. aeolicus,
and the mesophilic archeon M. voltae form hydrogen bonds with AMPbd. Also, the psychrophilic
B. globisporus forms only one hydrogen bound with an ATPlid residue. The hyperthermophilic archeon
S. acidocaldarius forms only two hydrogen bonds with residues which do not belong to AMPbd or
ADPbd—the substrates which were co-crystallyzed with 1NKS. Finally, B. stearothermophilus and
M. voltae interact with residues which do not belong to the AK active site.

Molecules 2018, 23, 1076
6 of 11
In Figure 2 multiple sequence alignment of nine sequences show that the residues involved
substrate binding are well conserved among AKs of all domain of life, with notable differences among
the archean enzymes where nine sequences were aligned.
Figure 2. Multiple sequence alignment of the AKs’ sequences. Marked in blue are the residues
belonging to the Walker A motif—phosphate-binding loop or P-loop; in the black background was
selected the consensus sequence [LIVMFYWCA]-LIVMFYW] (2) -D-G-[FYI]-P-R-X (3) -[NQ] and in
yellow were selected the residues strictly conserved in the adenylate kinase family; with pink were
selected the Lys and Val residues—well conserved in AKs. Multiple sequence alignment was performed
by the Clustal Omega program (https://www.ebi.ac.uk/Tools/msa/clustalo). * (asterisk) indicates
positions which have a single, fully conserved amino acid residue; : (colon) indicates conservation
between groups of strongly similar properties as following—roughly equivalent to scoring >0.5 in
the Gonnet PAM 250 matrix: STA, NEQK, NHQK, NDEQ, QHRK, MILV, MILF, HY, FYW; . (period)
indicates conservation between groups of weakly similar properties as following—roughly equivalent
to scoring =<0.5 and >0 in the Gonnet PAM 250 matrix: CSA, ATV, SAG, STNK, STPA, SGND, SNDEQK,
NDEQHK, NEQHRK, FVLIM, HFY
3. Discussion
The discovery of new antibacterial compounds is a great challenge because microorganisms are so
versatile that they inevitably develop smart resistance mechanisms. Before understanding the subtle
mechanisms of inhibitory activity of new compounds against an unusual target like AK, it is essential to

Molecules 2018, 23, 1076
7 of 11
accurately identify the targets. Otherwise, we will be stuck in time-consuming and costly experiments
without valuable results. AK is an interesting enzyme that still raises the interest of many research
teams even though its structure and kinetic properties are well known. When it comes to starting
a new project, such as synthesis and testing the inhibitory activity of new compounds, it is better to
use a well characterized target. In a previous work, we discovered a good inhibitory activity of some
thiazolidine derivatives against Escherichia coli [32] and Streptococcus pneumonia AKs [33]. Because the
derivatives 1a–f are soluble in dimethylformamide (DMF) the inhibitory activity of these compounds
against AK was possible only by performing by a 2,4-dinitrophenylhydrazine colorimetric assay,
not by the coupled spectrophotometric assay, a more reliable method for activity assay of AK [19].
Therefore, six 5-arylidene-2-thioxothiazolidin-4-one derivatives 1a–f were further explored to predict
their binding interactions by molecular docking. The results confirm that compound 1d C10H12BrNOS2
(R = 1 Br) inhibits bacterial AKs targeting the ATPlid residues of the active site. More, comparing the
cumulative docking data for AKs of bacterial and human origin, we could reject the null hypothesis
that there are no differences between the two groups (Table S1 and Table 2), therefore derivatives
1a–f are worth a deeper analyses of their inhibitory activity against bacterial AKs. Statistical tools are
useful as the first step of designing a study, providing valuable information about the relevance of
disparate data. After the preliminary statistical analysis, we continued the molecular docking study,
which is considered an empirical method, but providing some clues about interacting residues. During
conformational changes, water molecules removal from the active sites prevent nonproductive side
reactions, therefore, it is relevant to observe the influence of water molecules in inhibitory binding [26].
In the present study, we considered AKs as rigid receptors and the inhibitors as flexible ligands.
The molecular docking should be confirmed with experimental data, but our goal was first was first to
advance the analysis of the previous experimental data with an inexpensive method, then to improve
the structures of the best thiazolidine derivative for future studies.
For the derivatives 1a–f, we can compare the docking data with experimental data only for E. coli
and S. pneumoniae AKs, but these results are consistent and we can assume that the docking data
are representative for the rest of the tested AKs. The most concerning issue is about the inhibition
of the human AKs. Without experimental data, we can only draw a preliminary conclusion on their
interaction with derivatives 1a
study about inhibition of human AKs by derivatives 1a
low solubility of the 1a derivatives. Obviously, the colorimetric method previously used could
–f
. Our results are quite encouraging in designing a future experimental
–
f. Here is a major limitation due to the
–
f
be applied, but our results are promising for designing new thiazolidine compounds with similar
structures, but with a better solubility in Tris buffer in order to perform a more reliable kinetic study.
The main limitation of accurate kinetic studies of the inhibitory activity of the vast majority
of new synthetic compounds is their low water solubility, although promising results exist [19].
Thus molecular docking is an appropriate suitable alternative approach for deciphering the molecular
mechanisms of ligand binding, although there are limitations in the availability of 3D structures
in public databases. However, the present study demonstrates that AK is a valuable target for
certain thiazolidine derivatives and we will consider screening other enzymes and designing new
similar compounds.
Another interesting perspective is related to the pollution of the environment with synthetic
compounds and the way they can influence the species of microorganisms in certain ecologic niches.
Sulfur-containing compounds are found to be industrial waste which, in concert with other compounds,
influences environmental microorganisms. Perhaps only ecologists are interested in the dynamics
of environmental microorganisms, but the subject is not a minor one. Suffice it to say that the soil
environment directly influences the quality of drinking water. These problems are not difficult to
solve, but we are far from understanding them. We can only try to imagine what happens when some
synthetic compounds are spread in Nature. This is the main reason of including AKs of archeal origin
and AKs from bacterial species from extreme environments. AKs from the Archea and extremophile

Molecules 2018, 23, 1076
8 of 11
species interacts with AMPbd residues (Tables S2 and S3) contrary to their mesophylic counterparts,
which interact with ATP-binding residues.
Taken together, AK is an interesting enzyme not only for the development of new
thiazolidine derivatives or other sulfur-containing compounds but can play a crucial role in
understanding the sophisticated mechanisms developed by microorganisms or eukaryots in maintaining
energy homeostasis.
4. Materials and Methods
The methodology included: (1) ligand preparation for molecular docking; (2) acquisition of 3D
structures of AKs and preparing them as receptors for molecular docking; (3) molecular docking
protocol; and (4) statistical analysis.
4.1. Ligand Preparation
In the Figure 3 is represented the structure of derivatives 1a–f used as ligands in our protein
docking experiments and their chemical characteristics are given in Table S4. Their synthesis was
previously described in another paper [34]. Briefly they were obtained by Knoevenagel condensation
between various substituted aryl aldehydes and thiazolidin-4-on-2-thione, by refluxing in glacial
acetic acid, in the presence of anhydrous sodium acetate. The ChemDraw 10.0 program was used for
drawing, displaying and characterizing chemical structures. The optimized 3D structures were saved
as pdb files for molecular docking preparation by AutoDockTools.
1a (R = 2 OCH3)
1b (R = 3 Cl)
1c (R = 3 Br)
1d (R = 1 Br)
1e (R = 2 Cl)
1f (R = 1 OCH3)
Figure 3. The structure of the 5-arylidene-2-thioxothiazolidin-4-one derivatives 1a–f, R = various
substituents.
4.2. Adenylate Kinases Structure Acquisition and Preparation as Receptors
The main characteristics of the AKs included in the present study are presented in the Table S5.
In the present study we considered AKs from the three domains of life—Archaea, Bacteria, and
Eukarya [23,35–41]. For a better comparative analysis, AKs from the same species in closed
conformation—co-crystallized with Ap5A, AMP, or ATP—and in open conformation were first
collected. The enzymes were selected without mutations in their sequences. First, the AKs’ X-ray
3D structures were downloaded from the Protein Data Bank PDB (https://www.rcsb.org/structure).
From the 3D structure, all solvent molecules were removed and the substrates when proteins were
crystallized in closed conformation. For the multimeric AKs structures, we first chose the chain
crystallized with substrate or inhibitor then the chain crystallized without substrate or inhibitor.
When all chains are alike, the A chain was selected. The Dassault Systèmes BIOVIA—Discovery
Studio Modeling Environment (Release 2017, Dassault Systèmes, San Diego, CA, USA, 2016;
http://accelrys.com)—allows quick preparations of the protein structures. After removing water
molecules, and other substrates or inhibitors, the AKs were saved as pdb files for molecular docking

Molecules 2018, 23, 1076
9 of 11
preparation which involved the addition of non-polar hydrogen and of Kollman charges, computed
by AutoDockTools.
4.3. Molecular Docking Protocol
Although there are many programs for protein-ligand docking, we concluded that the
AutoDock4.2 program, distributed as open source under a GPL license, perfectly accomplished the
requirements for designing the present study (http://autodock.scripps.edu) [42,43]. For non-experts,
BIOVIA is a friendly program for further interpretations of docking data on Windows platform.
It allows the observation of the best docking conformation, permits one to visualize the protein-ligand
2D diagram of the interacting atoms, and offers protein-ligand complex 3D molecular visualization
tools. Ligand flexibility was assigned in the following steps—detect the root that minimizes the size
of the largest branch, set the number of torsions in the interval 1–6, then choose aromaticity criterion
at 7.50. Herein, we used an algorithm based on a rigid protein structure of AKs and the Lamarckian
Genetic Algorithm (LGA), with a maximum of 2,500,000 energy evaluations as searching parameter.
Otherwise, the search parameters and docking parameters were used as defaults. The Cygwin DLL
(cygwin1.dll) terminal, which provides functionality similar to a Linux distribution on Windows,
was necessary for running the molecular docking scripts.
4.4. Statistical Analysis
Data were expressed as mean
Student’s t-test—two-tailed hypothesis at 0.05 significance level. For comparing more than two samples,
±
standard deviation. The p-value was calculated with the
the One-Way ANOVA test and Kruskal–Wallis H test were performed.
Supplementary Materials: The following are available online: Figure S1: AKs in closed conformations docked
with derivative 1c; Figure S2: AKs in closed conformation docked with derivative 1d; Table S1: Comparison
of the binding energy ∆G (kcal/mol) and of the inhibition constant Ki (µM) of the closed-conformation AKs,
Table S2: The interactions of 1c with selected AKs; Table S3: The interactions of 1d with selected AKs; Table S4:
The characterization of 5-arylidene-2-thioxothiazolidin-4-one derivatives; Table S5: The main characteristics of the
AKs used in protein-ligand docking.
Author Contributions: M.I.I. conceived, designed and performed the molecular docking; analyzed
and interpreted the data; wrote the paper; O.O. conceived, designed and synthesis the 5-arylidene-2-
thioxothiazolidin-4-one derivatives.
Acknowledgments: The authors would like to thank the referees for courtly and useful comments.
Conflicts of Interest: The authors declare no conflict of interest.
References
1.
Achenbach-Richter, L.; Stetter, K.O.; Woese, C.R. A possible biochemical missing link among archaebacteria.
Nature 1987, 327, 348–349. [CrossRef] [PubMed]
2.
Brügger, K.; Chen, L.; Stark, M.; Zibat, A.; Redder, P.; Ruepp, A.; Awayez, M.; She, Q.; Garrett, R.A.;
Klenk, H.-P. The genome of Hyperthermus butylicus: A sulfur-reducing, peptide fermenting, neutrophilic
Crenarchaeote growing up to 108 degrees C. Archaea 2007, 2, 127–135. [CrossRef] [PubMed]
Sekiguchi, Y.; Ohashi, A.; Parks, D.H.; Yamauchi, T.; Tyson, G.W.; Hugenholtz, P. First genomic insights into
members of a candidate bacterial phylum responsible for wastewater bulking. PeerJ 2015, 3, e740. [CrossRef]
[PubMed]
3.
4.
5.
6.
7.
Barton, L.L.; Ritz, N.L.; Fauque, G.D.; Lin, H.C. Sulfur Cycling and the Intestinal Microbiome. Dig. Dis. Sci.
2017, 62, 2241–2257. [CrossRef] [PubMed]
Graham, J.S.; Schoneboom, B.A. Historical perspective on effects and treatment of sulfur mustard injuries.
Chem. Biol. Interact. 2013, 206, 512–522. [CrossRef] [PubMed]
Li, Y.; Geng, J.; Liu, Y.; Yu, S.; Zhao, G. Thiadiazole-a Promising Structure in Medicinal Chemistry.
ChemMedChem 2013, 8, 27–41. [CrossRef] [PubMed]
Schippa, S.; Conte, M.P. Dysbiotic events in gut microbiota: Impact on human health. Nutrients 2014, 6,
5786–5805. [CrossRef] [PubMed]

Molecules 2018, 23, 1076
10 of 11
8.
9.
Trotsko, N.; Przekora, A.; Zalewska, J.; Ginalska, G.; Paneth, A.; Wujec, M. Synthesis and in vitro
antiproliferative and antibacterial activity of new thiazolidine-2,4-dione derivatives. J. Enzyme Inhib.
Med. Chem. 2018, 33, 17–24. [CrossRef] [PubMed]
Garza, I.; Wallace, M.J.; Fernando, D.; Singh, A.; Lee, R.E.; Gerding, J.S.; Franklin, C.; Yendapally, R. Synthesis
and Evaluation of Thiazolidine Amide and N-Thiazolyl Amide Fluoroquinolone Derivatives. Arch. Pharm.
2017, 350. [CrossRef] [PubMed]
10. Mikaili, P.; Maadirad, S.; Moloudizargari, M.; Aghajanshakeri, S.; Sarahroodi, S. Therapeutic uses and
pharmacological properties of garlic, shallot, and their biologically active compounds. Iran. J. Basic Med. Sci.
2013, 16, 1031–1048. [PubMed]
11. Ahmed, S.; Zayed, M.; El-Messery, S.; Al-Agamy, M.; Abdel-Rahman, H. Design, Synthesis, Antimicrobial
Evaluation and Molecular Modeling Study of 1,2,4-Triazole-Based 4-Thiazolidinones. Molecules 2016, 21, 568.
[CrossRef] [PubMed]
12. Devi, P.B.; Samala, G.; Sridevi, J.P.; Saxena, S.; Alvala, M.; Salina, E.G.; Sriram, D.; Yogeeswari, P.
Structure-guided design of thiazolidine derivatives as Mycobacterium tuberculosis pantothenate synthetase
inhibitors. ChemMedChem 2014, 9, 2538–2547. [CrossRef] [PubMed]
13. Meyer, P.; Evrin, C.; Briozzo, P.; Joly, N.; Bârzu, O.; Gilles, A.-M. Structural and Functional Characterization
of Escherichia coli UMP Kinase in Complex with Its Allosteric Regulator GTP. J. Biol. Chem. 2008, 283,
36011–36018. [CrossRef] [PubMed]
14. Bucurenci, N.; Sakamoto, H.; Briozzo, P.; Palibroda, N.; Serina, L.; Sarfati, R.S.; Labesse, G.; Briand, G.;
Danchin, A.; Bărzu, O.; et al. CMP kinase from Escherichia coli is structurally related to other nucleoside
monophosphate kinases. J. Biol. Chem. 1996, 271, 2856–2862. [CrossRef] [PubMed]
15. Wiesmuller, L.; Noegel, A.A.; Barzu, O.; Gerisch, G.; Schleicher, M. cDNA-derived sequence of UMP-CMP
kinase from Dictyostelium discoideum and expression of the enzyme in Escherichia coli. J. Biol. Chem. 1990
265, 6339–6345. [PubMed]
,
16. Yamada, M.; Sugahara, M.; Hishitani, Y.; Nobumoto, M.; Nakazawa, A. Isolation and characterization of
mutated mitochondrial GTP:AMP phosphotransferase. J. Mol. Biol. 1998, 280, 551–558. [CrossRef] [PubMed]
17. Panayiotou, C.; Solaroli, N.; Johansson, M.; Karlsson, A. Evidence of an intact N-terminal translocation
sequence of human mitochondrial adenylate kinase 4. Int. J. Biochem. Cell Biol. 2010, 42, 62–69. [CrossRef]
[PubMed]
18. Liu, R.; Xu, H.; Wei, Z.; Wang, Y.; Lin, Y.; Gong, W. Crystal structure of human adenylate kinase 4 (L171P)
suggests the role of hinge region in protein domain motion. Biochem. Biophys. Res. Commun. 2008, 379, 92–97.
[CrossRef] [PubMed]
19. Saint Girons, I.; Gilles, A.M.; Margarita, D.; Michelson, S.; Monnot, M.; Fermandjian, S.; Danchin, A.; Barzu, O.
Structural and catalytic characteristics of Escherichia coli adenylate kinase. J. Biol. Chem. 1987, 262, 622–629.
[PubMed]
20. Berry, M.B.; Meador, B.; Bilderback, T.; Liang, P.; Glaser, M.; Phillips, G.N. The closed conformation of
a highly flexible protein: The structure of E. coli adenylate kinase with bound AMP and AMPPNP. Proteins
1994, 19, 183–198. [CrossRef] [PubMed]
21. Powers, S.G.; Griffith, O.W.; Meister, A. Inhibition of carbamyl phosphate synthetase by P1,
0
P5-di(adenosine 5 )-pentaphosphate: Evidence for two ATP binding sites. J. Biol. Chem. 1977, 252, 3558–3560.
[PubMed]
22. Rose, T.; Glaser, P.; Surewicz, W.K.; Mantsch, H.H.; Reinstein, J.; Le Blay, K.; Gilles, A.M.; Barzu, O. Structural
and functional consequences of amino acid substitutions in the second conserved loop of Escherichia coli
adenylate kinase. J. Biol. Chem. 1991, 266, 23654–23659. [PubMed]
23. Müller, C.W.; Schlauderer, G.J.; Reinstein, J.; Schulz, G.E. Adenylate kinase motions during catalysis:
An energetic counterweight balancing substrate binding. Structure 1996, 4, 147–156. [CrossRef]
24. Rundqvist, L.; Aden, J.; Sparrman, T.; Wallgren, M.; Olsson, U.; Wolf-Watz, M. Noncooperative folding of
subdomains in adenylate kinase. Biochemistry 2009, 48, 1911–1927. [CrossRef] [PubMed]
25. Wang, Y.; Makowski, L. Fine structure of conformational ensembles in adenylate kinase. Proteins 2018, 86,
332–343. [CrossRef] [PubMed]
26. Rogne, P.; Rosselin, M.; Grundstrom, C.; Hedberg, C.; Sauer, U.H.; Wolf-Watz, M. Molecular mechanism of
ATP versus GTP selectivity of adenylate kinase. Proc. Natl. Acad. Sci. USA 2018, 115, 3012–3017. [CrossRef]
[PubMed]

Molecules 2018, 23, 1076
11 of 11
27. Lee, Y.; Kim, J.W.; Lee, S.M.; Kim, H.J.; Lee, K.S.; Park, C.; Choe, I.S. Cloning and Expression of Human
Adenylate Kinase 2 Isozymes: Differential Expression of Adenylate Kinase 1 and 2 in Human Muscle Tissues.
J. Biochem. 1998, 123, 47–54. [CrossRef] [PubMed]
28. Panayiotou, C.; Solaroli, N.; Karlsson, A. The many isoforms of human adenylate kinases. Int. J. Biochem.
Cell Biol. 2014, 49, 75–83. [CrossRef] [PubMed]
29. Plant, T.; Journal, P. Identification of a Gene Encoding Adenylate Kinase Involved in Antifungal Activity
Expression of the Biocontrol Strain Burkholderia pyrrocinia CH-67. Plant Pathol. J. 2012, 28, 373–380.
30. Russell, P.J.; Conner, J.; Sisson, S. Sulfur specifically inhibits adenylate kinase in assays for creatine kinase.
Clin. Chem. 1984, 30, 1555–1557. [PubMed]
31. Russell, P.J.; Williams, A.; Avtla, D.; Chinn, E.; Taulane, J.P. Characteristics of Rabbit Muscle Adenylate
Kinase Inhibition by Sulfur and Recovery by Dithiothreitol. J. Enzyme Inhib. 1995, 9, 179–194. [CrossRef]
[PubMed]
32. Ionescu, M.I.; Radulescu, A.Z.; Banciu, H.L.; Ghiran, D. Inhibition of E.coli adenylate kinase by some
5-aryliden-2-thioxo-thiazolydin-4-one derivatives. Clujul Med. 2005, 3, 638–641.
33. Ionescu, M.I.; Costache, A.Z.; Oniga, O.; Banciu, H.L.; Lupan, I.; Lupan, I. Inhibition of Streptococcus
pneumoniae adenylate kinase by some 5-arylidene-thiazolidin-4-on-2-thione derivates. Rom. Rev. Lab. Med.
2013, 21, 93–99. [CrossRef]
34. Oniga, O.; Moldovan, C.; Muntean, A.; Verite, P.; Ionescu, M.; Tiperciuc, B. Synthesis and antibacterial
activity of some 5-aryliden-2-thioxo-thiazolydin-4-one. Rev. Med. Farm. 2004, 50, 183–186.
35. Schrank, T.P.; Bolen, D.W.; Hilser, V.J. Rational modulation of conformational fluctuations in adenylate kinase
reveals a local unfolding mechanism for allostery and functional adaptation in proteins. Proc. Natl. Acad.
Sci. USA 2009, 106, 16984–16989. [CrossRef] [PubMed]
36. Thach, T.T.; Luong, T.T.; Lee, S.; Rhee, D.K. Adenylate kinase from Streptococcus pneumoniae is essential for
growth through its catalytic activity. FEBS Open Bio 2014, 4, 672–682. [CrossRef] [PubMed]
37. Berry, M.B.; Phillips, G.N.J. Crystal structures of Bacillus stearothermophilus adenylate kinase with bound
Ap5A, Mg2+ Ap5A, and Mn2+ Ap5A reveal an intermediate lid position and six coordinate octahedral
geometry for bound Mg2+ and Mn2+. Proteins Struct. Funct. Genet. 1998, 32, 276–288. [CrossRef]
38. Bae, E.; Phillips, G.N. Structures and analysis of highly homologous psychrophilic, mesophilic, and
thermophilic adenylate kinases. J. Biol. Chem. 2004, 279, 28202–28208. [CrossRef] [PubMed]
39. Henzler-Wildman, K.A.; Thai, V.; Lei, M.; Ott, M.; Wolf-Watz, M.; Fenn, T.; Pozharski, E.; Wilson, M.A.;
Petsko, G.A.; Karplus, M.; et al. Intrinsic motions along an enzymatic reaction trajectory. Nature 2007, 450,
838–844. [CrossRef] [PubMed]
40. Criswell, A.R.; Bae, E.; Stec, B.; Konisky, J.; Phillips, G.N., Jr. Structures of Thermophilic and Mesophilic
Adenylate Kinases from the Genus Methanococcus. J. Mol. Biol. 2003, 330, 1087–1099. [CrossRef]
41. Vonrhein, C.; Bönisch, H.; Schäfer, G.; Schulz, G.E. The structure of a trimeric archaeal adenylate kinase.
J. Mol. Biol. 1998, 282, 167–179. [CrossRef] [PubMed]
42. Goodsell, D.S.; Olson, A.J. Automated docking of substrates to proteins by simulated annealing. Proteins
Struct. Funct. Genet. 1990, 8, 195–202. [CrossRef] [PubMed]
43. Morris, G.M.; Huey, R.; Lindstrom, W.; Sanner, M.F.; Belew, R.K.; Goodsell, D.S.; Olson, A.J. AutoDock4
and AutoDockTools4: Automated docking with selective receptor flexibility. J. Comput. Chem. 2009, 30,
2785–2791. [CrossRef] [PubMed]
©
2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).