Identification of potential inhibitor against CTX-M-3 and CTX-M-15 proteins: an in silico and in vitro study

Abstract

Extended-spectrum beta-lactamase (ESBL) producing Enterobacteriaceae infection is a serious global threat. ESBLs target 3rd generation cephalosporin antibiotics, the most commonly prescribed medicine for gram-negative bacterial infections. As bacteria are prone to develop resistance against market-available ESBL inhibitors, finding a novel and effective inhibitor has become mandatory. Among ESBL, the worldwide reported two enzymes, CTX-M-15 and CTX-M-3, are selected for the present study. CTX-M-3 protein was modeled, and two thousand phyto-compounds were virtually screened against both proteins. After filtering through docking and pharmacokinetic properties, four phyto-compounds (catechin gallate, silibinin, luteolin, uvaol) were further selected for intermolecular contact analysis and molecular dynamics (MD) simulation. MD trajectory analysis results were compared, revealing that both catechin gallate and silibinin had a stabilizing effect against both proteins. Silibinin having the lowest docking score, also displayed the lowest MIC (128 µg/mL) against the bacterial strains. Silibinin was also reported to have synergistic activity with cefotaxime and proved to have bactericidal effect. Nitrocefin assay confirmed that silibinin could inhibit beta-lactamase enzyme only in living cells, unlike clavulanic acid. Thus the present study validated the CTX-M inhibitory activity of silibinin both in silico and in vitro and suggested its promotion for further studies as a potential lead. The present study adopted a protocol through the culmination of bioinformatics and microbiological analyses, which will help future researchers identify more potential leads and design new effective drugs.
Communicated by Ramaswamy H. Sarma

Full text

Full Terms & Conditions of access and use can be found at
https://www.tandfonline.com/action/journalInformation?journalCode=tbsd20
Journal of Biomolecular Structure and Dynamics
ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/tbsd20
Identification of potential inhibitor against CTX-
M-3 and CTX-M-15 proteins: an in silico and invitro
study
Bipasa Kar, Chanakya Nath Kundu, Mahender Kumar Singh, Budheswar
Dehury, Sanghamitra Pati & Debdutta Bhattacharya
To cite this article: Bipasa Kar, Chanakya Nath Kundu, Mahender Kumar Singh, Budheswar
Dehury, Sanghamitra Pati & Debdutta Bhattacharya (2023): Identification of potential inhibitor
against CTX-M-3 and CTX-M-15 proteins: an in silico and invitro study, Journal of Biomolecular
Structure and Dynamics, DOI: 10.1080/07391102.2023.2192811
To link to this article: https://doi.org/10.1080/07391102.2023.2192811
View supplementary material
Published online: 30 Mar 2023.
Submit your article to this journal
View related articles
View Crossmark data

Identification of potential inhibitor against CTX-M-3 and CTX-M-15 proteins:
an in silico and in vitro study
Bipasa Kar
a,b
, Chanakya Nath Kundu
b
, Mahender Kumar Singh
c
, Budheswar Dehury
d
, Sanghamitra Pati
a
and
Debdutta Bhattacharya
a
a
Department of Health Research, Ministry of Health & Family Welfare, Govt. of India, ICMR-Regional Medical Research Centre, Bhubaneswar,
Odisha, India;
b
School of Biotechnology, Kalinga Institute of Industrial Technology, Bhubaneswar, Odisha, India;
c
Data Science Laboratory,
National Brain Research Centre, Gurgaon, Haryana, India;
d
Bioinformatics Division, ICMR-Regional Medical Research Centre, Nalco Square,
Bhubaneswar, Odisha, India
Communicated by Ramaswamy H. Sarma
ABSTRACT
Extended-spectrum beta-lactamase (ESBL) producing Enterobacteriaceae infection is a serious global
threat. ESBLs target 3rd generation cephalosporin antibiotics, the most commonly prescribed medicine
for gram-negative bacterial infections. As bacteria are prone to develop resistance against market-avail-
able ESBL inhibitors, finding a novel and effective inhibitor has become mandatory. Among ESBL, the
worldwide reported two enzymes, CTX-M-15 and CTX-M-3, are selected for the present study. CTX-M-3
protein was modeled, and two thousand phyto-compounds were virtually screened against both pro-
teins. After filtering through docking and pharmacokinetic properties, four phyto-compounds (catechin
gallate, silibinin, luteolin, uvaol) were further selected for intermolecular contact analysis and molecular
dynamics (MD) simulation. MD trajectory analysis results were compared, revealing that both catechin
gallate and silibinin had a stabilizing effect against both proteins. Silibinin having the lowest docking
score, also displayed the lowest MIC (128 mg/mL) against the bacterial strains. Silibinin was also
reported to have synergistic activity with cefotaxime and proved to have bactericidal effect. Nitrocefin
assay confirmed that silibinin could inhibit beta-lactamase enzyme only in living cells, unlike clavulanic
acid. Thus the present study validated the CTX-M inhibitory activity of silibinin both in silico and
in vitro and suggested its promotion for further studies as a potential lead. The present study adopted
a protocol through the culmination of bioinformatics and microbiological analyses, which will help
future researchers identify more potential leads and design new effective drugs.
ARTICLE HISTORY
Received 19 January 2023
Accepted 10 March 2023
KEYWORDS
Virtual screening; ESBL;
Phyto-compounds; molecu-
lar dynamic simulation
1. Introduction
Antibiotic resistance is an emerging threat to public health.
Due to antibiotic-resistant bacterial infections, directly 1.27
million and associated 4.95 million people died in 2019 glo-
bally (University of Oxford, 2022). It was considered one of the
top ten global public health problems (EClinicalMedicine,
2021). Recently, antibiotic-resistant gram-negative bacteria
have been on the rise globally. Two-thirds of the deaths
caused by multi-drug-resistant bacterial infections happen due
to gram-negative bacterial infections (Fair & Tor, 2014;U.S.
Department of Health and Human Services & CDC, 2019).
Beta-lactam antibiotics are generally prescribed to treat gram-
negative bacterial infections (Pandey & Cascella, 2022). Because
of the increased irresponsible use of beta-lactams (primarily 3rd
generation cephalosporins), bacteria developed resistance against
them. Gram-negative bacteria rely on producing beta-lactamase
enzyme, which destroys the antibiotic’s beta-lactam ring and
prevents from action (Ur Rahman et al., 2018). Among beta-
lactamases, extended-spectrum beta-lactamase-producing (ESBL)
Enterobacteriaceae is considered a severe threat by the World
Health Organization (WHO) and Center for Diseases Control and
Prevention (CDC) (U.S. Department of Health and Human
Services & CDC, 2019; WHO, 2019). ESBLs recognize all groups
of beta-lactams except carbapenem. They mainly target broad-
spectrum 3rd generation cephalosporins (Dhillon & Clark, 2012).
ESBL genes are highly transmissible and are often associated
with other drug-resistant genes, making it difficult for medical
practitioners to treat patients (Castanheira et al., 2021). Over the
last half-decade, only in the USA, a more than 50% rise in ESBL-
producing Enterobacteriaceae infections was recorded among
hospitalized patients (Kadri, 2020). In India, several reports indi-
cate the prevalence of ESBL producers is between 60-80%
(Govindaswamy et al., 2019).
Among ESBL producers, the CTX-M variant is the most
prevalent. CTX-M beta-lactamase enzyme can be divided into
six sub-lineages (CTX-M-1, CTX-M-2, CTX-M-8, CTX-M-9, CTX-
M-25, CTX-M-45) based on the similarity on the amino acid
level. Among six sub-lineages, CTX-M-1 group members are
increasingly reported worldwide (Bevan et al., 2017; Rossolini
et al., 2008). One of the variants in the CTX-M-1 sub-lineage,
CTX-M-15, is most commonly reported globally. Since their
CONTACT Bipasa Kar karbipasa@gmail.com
Supplemental data for this article can be accessed online at https://doi.org/10.1080/07391102.2023.2192811.
ß2023 Informa UK Limited, trading as Taylor & Francis Group
JOURNAL OF BIOMOLECULAR STRUCTURE AND DYNAMICS
https://doi.org/10.1080/07391102.2023.2192811

first outbreak in 2004, they outpaced other variants (Doi
et al., 2017). Another CTX-M variant, CTX-M-3, which is also
increasingly reported from different parts of the world,
belongs to the CTX-M-1 group (Dutour et al., 2002; Ramdani-
Bouguessa et al., 2006; Zeynudin et al., 2018). This increasing
prevalence of CTX-M caused resistance demands to find an
effective inhibitor for CTX-M enzymes.
In search of the treatment of CTX-M-producing bacterial
infections, medical practitioners commonly use carbapenem
or beta-lactam antibiotics in combination with a beta-lacta-
mase inhibitor (Peirano & Pitout, 2019). But increased carbape-
nem usage results in increased carbapenem resistance, against
which no alternative treatments are available. However,
increased use of beta-lactamase inhibitors results in beta-lacta-
mase inhibitor resistance in bacteria. The ESBL inhibitors are
often comprised of the same structural feature as beta-lactam
antibiotics, which increases the probability of developing
resistance against them (Nichols et al., 2012). So to treat CTX-
M-producing bacterial infections, new antibiotics must be dis-
covered. Discovering a new antibiotic class is challenging,
time-taking, and highly pricey. In contrast, bacteria quickly
develop resistance against newly exposed antibiotics, decreas-
ing the market shelf life of the antibiotics (Hutchings et al.,
2019). Thus the pipeline of new antibiotics is drying, and there
is an urgent need to find effective anti-microbials for fighting
the upcoming drug resistance.
In the current scenario, plants can serve as a potential
alternative. Traditional people have relied on plant extracts
to treat infectious diseases; today, 80% of the world popula-
tion still relies on conventional medicine to cure and boost
immunity. Earlier plant extracts and plant-based powder
were commonly used in traditional medicine. Still, gradually,
due to the advancement of modern technologies of phyto-
compounds isolation, they are considered to be tested for
their medicinal properties. As phyto-compounds are com-
paratively easier to isolate than their synthetic counterparts
and cost-friendly, they are fascinating among researchers
(Khameneh et al., 2019). Until now, only 1% of plants are
characterized by their secondary metabolites, so it’s been
considered a highly undervalued resource by World Health
Organization (WHO) and needs more exploration. There are
several ongoing studies on using plant extracts as beta-lacta-
mase inhibitors, but very few focused on the detailed mech-
anism of action (Anand et al., 2019).
In the present study, we focused on identifying phyto-
compound that serve as CTX-M-3 and CTX-M-15 inhibitors.
To screen phyto-compounds with potential CTX-M inhibitory
activity, we employ the combination of both in silico and
in vitro studies. Virtual screening of phyto-compounds
against both proteins helps to select the best potential com-
pounds. We used molecular docking and pharmacokinetic
properties screening to identify the best potential phyto-
compounds based on their docking score and the probability
of acting best inside the body. Next, we identified the four
best compounds (catechin gallate, luteolin, uvaol, and silibi-
nin) and performed molecular dynamic simulations. The
simulation results of complexes with phyto-compounds were
compared with complexes with antibiotics and protein alone.
Next, the phyto-compounds were subjected to in vitro study,
where minimum inhibitory concentration (MIC) and min-
imum bactericidal concentration (MBC) were calculated. Next,
the synergistic activity of the phyto-compounds with cefotax-
ime was diagnosed using checker board assay. The beta-lac-
tamase inhibitory activity of the best-selected compound
was tested using nitrocefin assay.
2. Methodology
To start the structure-based virtual screening process, we
have selected the receptor protein (CTX-M-15) structure from
the RCSB protein data bank (PDB) (ID: 4HBT). The crystallo-
graphic structure of CTX-M-3 protein was not available in
PDB, so we modeled the structure and performed docking.
The list of phyto-compounds was retrieved from Indian
Medicinal Plants, Phyto-chemistry, and Therapeutics (IMPPAT)
database (Mohanraj et al., 2018). The structures of the phyto-
compounds were downloaded from the PubChem database.
2.1. Hardware and software specification
All computations were performed in Intel core i7-4770 proces-
sor@3.4 GHz, 64GB RAM, 12threads, with 2 G.B. graphics card
of NVIDIA Quadro K2200 running Ubuntu 18.04LTS as the
operating system. For molecular docking, AutoDock Vina soft-
ware and molecular dynamic simulation GROMACS 2018.7 ver-
sion was used (Pronk et al., 2013; Trott & Olson, 2009).
2.2. Homology modeling of CTX-M-3 protein
For structure-based virtual screening, both protein and lig-
and structures are required. But unavailability of the CTX-M-3
crystal structure in the RCSB protein data bank (PDB) com-
pelled us to model the structure. To perform homology mod-
eling of CTX-M-3 protein, the amino acid sequence was
downloaded from the UniProtKB (Universal Protein Resource
Knowledgebase) database (http://www.uniprot.org/) with
accession no Q7X573, which consists of 291Aa. To match the
target sequence, the SWISS-MODEL template library was
searched with BLAST and HHBlits (Arnold et al., 2006;
Camacho et al., 2009; Remmert et al., 2012). After the target-
template alignment, the CTX-M-15 structure complexed with
FPI-1523 was selected as the top hit with 99.53% similarity.
By using the ProMod3 modeling engine, a new model was
built. The quality of modeled 3D structure was assessed
using Qualitative Model Energy Analysis (QMEAN) and Global
Model Quality Estimation (GMQE) scores. The structure is
considered poor quality if the QMEAN score is below 4.0.
For GMQE, the score ranges from 0 to 1, and the higher
value represents high reliability (Benkert et al., 2011). Next,
the Ramachandran plot of the modeled structure was gener-
ated from the online server, and the protein structure was
further verified in Structural Analysis and Verification Server
(SAVES) (http://nihserver.mbi.ucla.edu/SAVES). In the SAVES
server, the PROCHECK program was used to understand the
stereo-chemical quality of the protein structure, and ERRAT
was used to assess the quality of the 3D model (Colovos &
2 B. KAR ET AL.

Yeates, 1993; Laskowski et al., 1993). The modeled structure
was used for molecular docking.
2.3. Target protein and ligand preparation
The 3D X-ray crystallographic structure of CTX-M-15 protein
was downloaded from RCSB PDB with PDB ID:4HBT. The
structure had low resolution (1.10 Å) and no mutation (Lahiri
et al., 2013). After checking through the structure with
SwissPDBViewer, they were prepared for molecular docking
using AutoDock tools (Morris et al., 2009). Water molecules
and heteroatoms were removed from the structure. Polar H
atoms were added to the structure. The Kollman charge was
added, and the Gasteiger charge was computed. Then the
.pdbqt format of the protein structure was saved for molecu-
lar docking. Both the modeled and downloaded protein
structures were prepared through the same protocol.
The retrieved phyto-compounds from the IMPPAT data-
base were downloaded from the PubChem database. The
structures were geometry optimized using Avogadro soft-
ware (Hanwell et al., 2012). The structures were saved in their
.pdbqt format using AutoDock tools for further analysis.
2.4. Molecular docking-based virtual screening
Molecular docking was performed to predict the ligands’
binding affinity and ideal binding pose against the target
protein. The knowledge about active site residues was col-
lected from previous literature studies Lahiri et al., 2013).
Receptor grids (40X40X40 Å
3
) were generated in AutoDock
tools 1.5.6, covering active site residues in an enclosed box.
AutoDock Vina was used for molecular docking with exhaust-
iveness set to 50. Lamarckian genetic algorithm of 100,000
energy evaluations for each run with a maximum of 27,000
generations with a population of 100 individuals was set for
each docking. Against both CTX-M-3 and CTX-M-15 proteins,
two thousand phyto-compounds were docked individually,
and their docking scores were compared with cefotaxime.
After docking, the best two-fifty compounds were
selected, comparing the docking score with the reference
drug cefotaxime. The phyto-compounds were screened for
their pharmacokinetic properties using pkCSM and
SwissADME online servers (Daina et al., 2017; Pires et al.,
2015). Pharmacokinetic properties like gastrointestinal
absorption (%), blood-brain barrier (BBB) permeability, AMES
toxicity (mutagenicity), hepatotoxicity, oral rat acute toxicity
lethal dose (LD50), and the number of violations in Lipinski’s
rule of five(RO5) were investigated for filtering potential
phyto-compounds (Lipinski et al., 2012). Subsequently, the
compounds were tested for their behavior as ‘Pan Assay
Interference Compounds’(PAINS). PAINS compounds react
nonspecifically with a broad range of targets, increasing the
chance of false positives (Baell & Walters, 2014). Compounds
with gastrointestinal absorption (>50%), no BBB permeability,
zero to one violation in RO5, no toxicity, LD50>2.0 mol/kg,
and in PAINS, zero to one alert were considered for filtering
the potential leads. Four phyto-compounds (Catechin Gallate,
Uvaol, Luteolin, and Silibinin) were selected based on low
docking scores and favorable pharmacokinetic properties for
molecular dynamics simulation. Each docked complex struc-
ture was visualized using PyMol (www.pymol.org) and Biovia
Discovery Studio Visualizer. The molecular interactions were
analyzed using LigPlot þv1.4.5 (Laskowski & Swindells, 2011).
The selected phyto-compounds were re-docked against the
target proteins under the same conditions to confirm their
docking result.
2.5. Molecular dynamics (MD) simulation
The ten complex structures (protein in complex with anti-
biotic cefotaxime and four phyto-compounds) and two apo-
proteins (CTX-M-3 and CTX-M-15) were selected for the
200 ns MD run. Each MD run was performed in the
GROMACS 2018.7version mmpbsa (Berendsen et al., 1995).
The ligand topologies were generated using the CHARMM
General force field (Vanommeslaeghe et al., 2012). The sys-
tems were simulated using an all-atom charmm36-forcefield in
a cubic box with a TIP3P water model (Huang & MacKerell,
2013). Both systems (apo&holo) were neutralized by adding
counter ions (Na
þ
/Cl
-
) at 0.15 M strength. Before the simula-
tion, all the systems were subjected to 100,000 steps of
energy minimization using the steepest gradient algorithm to
remove any geometric strain. After energy minimization, all
the systems were subjected to two consecutive equilibrations
followed by production runs using the leapfrog algorithm and
Berendsen Coupling (Berendsen et al., 1984). The first equili-
bration was carried during 5 ns at 313 K in NVT, followed by
10 ns in NPT at 1.0 bar ensembles. LINCS algorithm and par-
ticle-mesh Ewald (PME) algorithm were used to regulate the
covalent bonds and long-range ionic interactions (Ewald,
1921;Hessetal.,1997). The production simulation runs were
set to 200ns, and every coordinate was saved at every 0.2ps.
The edge effects of all periodic boundary conditions were cor-
rected in all trajectories. After the simulation run, the trajecto-
ries were analyzed using GROMACS utility toolkits. The
stability of each system was analyzed by calculating the back-
bone root mean square deviation (RMSD), the radius of gyr-
ation (Rg), and Ca-root mean square fluctuation (RMSF).
Hydrogen bonds (H bond) between protein and ligands were
calculated. The generated 2D graphs and structures can be
visualized using XmGrace, and PyMol, respectively.
2.6. Principal component and dynamic cross-correlation
analysis
The trajectory dynamics complexity data were analyzed with-
out ignoring the variability through the principal component
analysis (PCA) (Chen et al., 2021,2022; Dehury et al., 2014).
Through principal component analysis, a covariance matrix
containing eigenvectors was generated. Each eigenvector
represents the correlated motion of the protein. Each eigen-
value corresponding to each eigenvector represents the
amplitude of each eigenvector, which further helps to under-
stand the atomic contribution of motion in each system. 2D
trajectory projection (P.C.) was prepared by overlapping the
first two eigenvectors, where higher dispersion represents
JOURNAL OF BIOMOLECULAR STRUCTURE AND DYNAMICS 3

higher flexibility. The built-in tools of GROMACS, including
gmx covar and gmx anaeig, were used to perform the PCA.
The dynamic cross-correlation map (DCCM) was per-
formed to check the correlated motion of residues through-
out the MD run. The negative value (pink) represents
anti-correlation, signifying the residue’s movement in the
opposite direction. The positive value represents the move-
ment of the residues in the same direction. Bio3D package in
the R studio environment was used to calculate DCCM
(Grant et al., 2006). From the last 100 ns, 2000 frames were
selected, and Ca-atoms were considered for the analysis.
2.7. Binding free energy analysis using MM/PBSA
Binding free energy calculation plays a significant role in
evaluating the dynamic interaction between protein and lig-
and. The binding free energy was calculated for each com-
plex structure using the g_mmpbsa tool in GROMACS
(Kumari et al., 2014). From the last 100 ns, 2000 frames were
selected for the analysis. The solvation properties of the
complex structures during MD were analyzed using the
g_mmpbsa tool. Two contributions to the free energy were
considered for the analysis: (E
MM
) one included both bonded
(bond angle, torsion energies) and nonbonded (vdW, electro-
static) interactions, and the second one included Esol, describ-
ing the sum of the two energy terms polar and non-polar
solvation energy in an implicit water model. The binding free
energy calculation does not include entropic contribution
(E. Wang et al., 2019).
2.8. Identification of bacterial strains with CTX-M-3 and
CTX-M-15 genes
Bacterial isolates were collected from clinical samples of urin-
ary tract infections. Pure bacteria cultures were maintained
in a Nutrient Agar (N.A.) stab culture. Bacterial strains were
identified with the help of biochemical tests and an auto-
matic bacterial identification system (Vitek 2 Compact,
bioM
erieux). Antibiotic susceptibility tests were performed
using CLSI guidelines through the disc diffusion method
(CLSI, 2019). Bacterial strains resistant to cefotaxime (CTX)
and ceftazidime (CAZ) and susceptible to imipenem (IPM)
were selected for further ESBL screening tests through the
double-disc diffusion method. The ESBL-positive tested iso-
lates were selected for bacterial genomic DNA isolation
through the phenol-chloroform-isoamyl alcohol. Bacterial
DNA was stored at 20 C for long-term usage. Bacterial
DNA was amplified using polymerase chain reaction (PCR)
with the primers CTX-M-3-F (AATCACTGCGTCAGTTCAC), CTX-
M-3-R (TTTATCCCCCACAACCCAG), CTX-M-15-F (AGAATAAGG
AATCCCATGGTT) and CTX-M-15-R (ACCGTCGGTGACGAT
TTTAG) ((Kar et al., 2021; Mendonc¸a et al., 2007). The anneal-
ing temperatures for CTX-M-3 and CTX-M-15 were set at
50

C and 52

C, respectively. The positive tested isolates were
selected for the present study.
2.9. Determination of minimum inhibitory concentration
(MIC) and minimum bactericidal
concentration (MBC)
Phyto-compounds (catechin gallate, uvaol, luteolin, and silibi-
nin) and antibiotics (cefotaxime, gentamicin, meropenem)
were purchased from Sigma Aldrich. Both minimum inhibitory
and minimum bactericidal concentrations were calculated to
check the in vitro antibacterial activity of the phyto-com-
pounds. Minimum inhibitory concentration (MIC) is calculated
as the lowest concentration of an antimicrobial that inhibits
visible bacterial growth. Minimum bactericidal concentration
(MBC) is the minimum concentration of the antibacterial, kill-
ing 99.9% of the initial bacterial inoculum. The ratio of
MBC/MIC helps to determine if the phyto-compound is bacter-
iocidal (4) or bacteriostatic (>4) (Mogana et al., 2020). A
resazurin microplate assay detected visible bacterial growth in
microtiter plates .
CTX-M-3 and CTX-M-15 positive E.coli cells were grown to
0.5 McFarland standards (10
8
CFU/mL) in Muller Hinton Broth.
Anti-microbials were distributed through the serial dilution
technique in a 96-well microtiter plate, and the total volume
was 200 mL. Meropenem and 1xPBS were used as the positive
and negative control, respectively. At 37 C, sealed microtiter
plates were incubated overnight. The viability of cells can be
checked by adding 6 mL of 0.3% resazurin to each well and
then incubating for 3 hrs. After calculating the MIC of the
phyto-compounds, the microtiter plate wells showed no bac-
terial growth and were further streaked into nutrient agar
plates. After overnight incubation, the bacterial colonies
(Colony Forming Units: CFU) were counted. Each experiment
was replicated in triplicates for confirmed results.
2.10. Checkerboard assay
Checkerboard synergy testing is a way of checking bacterial
growth in combination with phyto-compounds and antibiot-
ics. CTX-M-3 and CTX-M-15 positive E.coli cells were inocu-
lated in nutrient broth and incubated for 16 hrs at 37

C. In a
96-well microtiter plate, the last two columns were kept as a
control for E.coli growth and plate sterility. Muller Hinton
Broth (100 mL) was added to each well. Phyto-compounds
ranged from 1024 to 8 mg/mL, and antibiotics (CTX) from
256 mg/mL to 0.5 mg/mL were distributed in a 96-well micro-
titer plate in a two-fold serial dilution so that each well con-
tained a unique combination. The final bacterial inoculum
(5 10
5
CFU/mL) was added to each well except for the
negative control. The plates were incubated at 37 C over-
night. Synergy was calculated in fractional inhibitory concen-
tration (RFIC index). The equation for calculating RFIC index
is the following: RFIC ¼FIC A þFIC B, where FIC A is the MIC
of the phyto-compound in combination with antibiotic/MIC
of the phyto-compound alone and FIC B is the MIC of the
antibiotic in combination with phyto-compound/MIC of the
antibiotic alone. The interaction between phyto-compound
and antibiotic was estimated in terms of RFIC index, such as
synergistic (RFIC <0.5), additive (0.5 RFIC <1), indifferent
(1 >RFIC <4), antagonist (RFIC >4) effect.
4 B. KAR ET AL.

2.11. Nitrocefin assay
Nitrocefin is a yellow-colored cephalosporin, and with the
reduction through the beta-lactamase enzyme, it changes its
color from yellow to red, which can be visible at 490 nm.
After checking the anti-ESBL activity of the phyto-com-
pounds, it is necessary to identify if the phyto-compounds
are inhibiting the beta-lactamase enzyme or utilizing any
other secondary mechanism to prevent bacterial growth. A
new bacterial suspension (OD
620
nm¼0.1) was created from
the overnight grown bacterial culture and incubated at room
temperature for 2 hrs. After the bacteria reached exponential
growth (OD
620
nm¼0.5), 2 ml of bacterial aliquots were incu-
bated in silibinin and gentamicin at MIC/2 and MIC/4 con-
centration and kept for incubation at 37 C for 10 minutes.
To verify the beta-lactamase inhibitory activity, 0.05% nitro-
cefin solution was added, and the change of color was
recorded in an ELISA microplate reader at 480 nm (Ohene-
Agyei et al., 2014; Rahman & Khan, 2019). Gentamicin and
beta-lactamase inhibitor clavulanic acid was used as the
negative and positive control, respectively. To check if the
anti-microbials have direct beta-lactamase inhibitory activity,
the bacterial lysate was prepared using centrifugation and
followed the same protocol. All the experiments were per-
formed in triplicates to confirm the outcome.
3. Result and discussion
The threat associated with emerging antibiotic resistance has
been multiplied due to the sudden appearance of the
COVID-19 pandemic. Due to the COVID-19 pandemic, antibi-
otics’unnecessary consumption has increased, which raises
the antibiotic resistance burden (Lai et al., 2021). WHO
declared that antibiotic resistance is an upcoming pandemic
and CTX-M producing bacterial spread, decreasing the effect-
ivity of broad-spectrum 3rd generation cephalosporins
(Nadimpalli et al., 2021). Considering the urgent need for
new antibacterial and pharmacy companies’reduced interest
in investing in developing new antibiotics forces researchers
to find alternatives. Plants, with their novel secondary metab-
olite structures, are considered to be potential antimicrobials.
Several studies aim to find medicinal plants that treat beta-
lactam-resistant bacterial infections, which can further help
us fight drug resistance (Liu et al., 2018; Toudji et al., 2017;
Wang et al., 2020). Based on the previous literature, the
study selected two thousand phyto-compounds from thirty-
six medicinal plants. The selected individuals were tested for
their effectivity against CTX-M-3 and CTX-M-15 proteins.
3.1. Homology modeling and structural validation
To prepare the 3D modeled structure of CTX-M-3 protein, we
used SWISS-MODEL, where multiple sequence alignment
revealed 99.53% similarity of CTX-M-15 protein with CTX-M-3.
During template library search, PDB ID: 5fa7.1.A was used as
the template. The GMQE score of the modeled structure was
0.98, and QMEAN Z-Score was 0.76, confirming the modeled
structure’s reliability. Ramachandran plots revealed that
98.12% of amino acids lie in the favorable regions, and the
MolProbity score is 0.55. In Ramachandran plots, the green
color represents favored regions. There is only 0.47%
Ramachandran outliers region. The Ramachandran plot of
the modeled structure is depicted in Supplementary Figure S1.
The lower MolProbity score represents a good structure (V. B.
Chen et al., 2010). After building the reliable model, the struc-
ture was further validated in the SAVES server. ERRAT is an
overall quality factor that describes nonbonded atomic interac-
tions; a higher value represents higher quality. The ERRAT
valueforthemodeledstructureis98.05,whereastheERRAT
value >50 represents higher quality. PROCHECK determines
the stereochemical quality of the protein structure by analyzing
the overall structural geometry. While calculating PROCHECK,
the modeled structure revealed no warnings, and the modeled
structure was used for molecular docking.
3.2. Molecular docking-based virtual screening
Among two thousand phyto-compounds, structure-based vir-
tual screening helps narrow the search process of effective
ones. Both protein and ligand structures are known in molecu-
lar docking, and AutoDock Vina was used to performing
molecular docking (Trott & Olson, 2009). Molecular docking
helps to give an idea about the binding affinity and binding
position of the ligands against the target protein. In molecular
docking, the docking scores of the phyto-compounds were
compared with cefotaxime. The lower docking score value rep-
resents more binding affinity with the target protein. Then the
best two hundred and fifty compounds (comparing the docking
scores with reference drug cefotaxime) were selected for further
pharmacokinetic parameter screening. Molecular docking fol-
lowed by predicting the drug’s reaction after entering the body
help to select the best potential leads. The docking score of
cefotaxime against CTX-M-3 protein and CTX-M-15 protein
were 6.2 kcal/mol and -6kcal/mol, respectively. The phyto-
compounds were selected having docking scores lower than
the reference cefotaxime for ADMET (absorption, distribution,
metabolism, excretion, and toxicity) properties screening.
Screening ADMET properties increases the probability of effi-
cient lead selection and is crucial in the early stages of drug dis-
covery (Guan et al., 2019).
After screening the pharmacokinetic parameters of two
hundred and fifty compounds, four phyto-compounds (cat-
echin gallate, luteolin, silibinin, uvaol) were selected as
potential leads. The list of best two hundred and fifty com-
pounds with their docking scores and pharmacokinetic
parameters were listed in the Supplementary Table S1.
Among the best four phyto-compounds, silibinin was
reported to have the highest binding affinity based on the
lowest docking score (-8.8 kcal/mol, 9.0 kcal/mol), followed
by catechin gallate and two other compounds, luteolin, and
uvaol. During ADMET properties screening, it is visible that
none of them has toxicity and mutagenic properties (AMES).
Oral rat acute toxicity, determined by lethal dose 50 (LD50),
signifies a lower value to be more toxic. High gastrointestinal
absorption and non-blood brain barrier (BBB) permeability
are essential for choosing leads. It was stated in Lipinski’s
JOURNAL OF BIOMOLECULAR STRUCTURE AND DYNAMICS 5

Rule of Five (RO5) that drugs with more than two or three
violations will not be suitable as orally active drugs, so com-
pounds with one or no violations were selected for further
study (Doak et al., 2014).
The synthetic accessibility (SA) score is another important
criterion for medicinal chemists to predict the difficulty syn-
thesizing the compound for mass production. It can be cal-
culated using the knowledge of the compound’s synthesis
Figure 1. Non-bonding interactions of the complex structures. (Yellow dot lines represent H bonds). The ligands are shown in green color in thick stick format. The
interacting amino acids were shown in thin stick format. The images were generated using Pymol. The structures of the ligands (K) (CG: Catechin Gallate; Lut:
Luteolin; Sil: Silibinin; Uva: Uvaol; CTX: Cefotaxime).
6 B. KAR ET AL.

and molecular complexity. A higher SA score represents
more difficulty in synthesis, and a lower S.A. value represents
the ease of the synthesis process (Ertl & Schuffenhauer,
2009). Among the four phyto-compounds, uvaol has the
highest SA score, which depicts its difficulty in synthesis in
an artificial environment. PAINS compounds confuse medi-
cinal chemists by showing false-positive results, but they
never showed their effectiveness in bioassay. More alerts
describe the chances of less bioactivity (J. Baell & Walters,
2014). In the present study, catechin gallate showed one
alert for catechol and luteolin for catechol A, but the rest
two were no PAINS compounds(J. B. Baell, 2016). The com-
pounds with two or more PAINS alerts were rejected for this
study.
From this structure-based virtual screening, the selected
four phyto-compounds (CG, luteolin, uvaol, and silibinin) and
their docking scores with respective ADMET properties are
listed in Table 1.
3.3. Intermolecular contact analysis
The next step after identifying four potential phyto-com-
pounds is analyzing intermolecular contacts in the com-
plexes. Hydrogen bonding and hydrophobic interactions play
a significant role in providing structural stability in any pro-
tein-ligand complex. The docked complex structures were
visualized through PyMol, and the interactions were con-
firmed through LigPlot þv1.4.5. The four phyto-compounds
in complex with CTX-M-3 and CTX-M-15 proteins were com-
pared with the reference drug cefotaxime. CTX-M-15 protein
has two structural domains, all aand an a/bdomain, where
the active site was located between the groove of these two
domains. CTX-M-15 protein differs from CTX-M-3 protein by
only one amino acid(G240D) (Cant
on et al., 2012). CTX-M
proteins are called serine beta-lactamase because of using
serine as a catalytic residue (Peirano & Pitout, 2019). The
detailed intermolecular contact analysis revealed the possible
binding pockets in the target protein.
Due to small-sized, geometry-optimized ligands used for
molecular docking, hydrophobic interactions are considered
influential in stabilizing protein-ligand interaction compared
to satisfying geometric constraints associated with electro-
static interactions (Ferreira De Freitas & Schapira, 2017).
According to the docking scores, silibinin had the lowest
docking score (i.e. highest binding affinity) against both pro-
teins. While interacting with CTX-M-3 protein, there are only
five H bonds (Ser105, Thr191, Ser212, Asp214, Lys244) and
fifteen hydrophobic interactions (Lys48, Asn79, Tyr80, Tyr104,
Asn107, Asn145, Asn189, Thr191, Lys209, Thr210, Ser212,
Lys244, Ala245, Glu246, Ser247) involved. Catechin gallate
(CG) had a docking score 8.0 kcal/mol against CTX-M-3 pro-
tein and four H bonds (Ser45, Asn79, Asn107, Asp214) and
eleven hydrophobic interactions (Tyr80, Glu141, Asn145,
Thr146, Lys209, Tyr215, Thr217, Gln242, Pro243, Ser247,
Arg249) were found to have to stabilize effect against the
target protein. Uvaol has the highest docking score among
the four compounds mentioned to have zero H bonds with
CTX-M-3 protein.
For CTX-M-15 protein, silibinin had the lowest docking
score (-9.0 kcal/mol), and four H bonds (Ser70, Asn132,
Gly241, Pro268) were involved in the complex formation.
There were five H bonds (Ser70, Ser130, Asn132, Asn170,
Ala270) and eight hydrophobic interactions (Asn104, Tyr105,
Glu166, Pro167, Thr171, Gly238, Gly239, Lys269) were
involved in the complex formation of CTX-M-15 protein with
CG (Figure 1). The reference drug cefotaxime had formed
five H bonds with (Ser130, Asn132, Ser237, Pro268, Ala270)
CTX-M-15 and (Ser45, Ser105, Asn107, Ser212, Pro243) CTX-
M-3 protein. The structures of phyto-compounds and anti-
biotic cefotaxime are also depicted in Figure 1K.
Superimposed structures of complexes with both CTX-M-3
and CTX-M-15 proteins revealed the binding pocket consists
of beta-sheets and a helical loop (Supplementary Figure S2).
The active site residues of CTX-M-3 protein are Ser45, Asn79,
Tyr80, Ser105, Asn107, Thr191, and Asp214. The active site
residues of CTX-M-15 protein can be listed as Ser70, Asn104,
Tyr105, Ser130, Asn132, Asn170, Thr235, Ser237, Gly238,
Gly239, Gly241, Pro268, and Ala270.
While understanding the intermolecular interaction of the
complex structures, it can be concluded that the binding
pockets of both phyto-compounds and reference drug cefo-
taxime are the same. Thus the ligands and cefotaxime com-
pete for the binding site while binding with the protein, so
we can conclude that the phyto-compounds are involved in
a competitive mode of inhibition.
3.4. Analysis of molecular dynamics (MD) simulation
MD simulation is a computational approach to recognizing phys-
ical interactions in a biophysical environment. In the present
study, CTX-M-3 and CTX-M-15, both proteins alone (apo-protein)
and in combination with four phyto-compounds (CG, luteolin,
uvaol, silibinin) and one antibiotic cefotaxime (CTX) were sub-
jected to 200 ns MD simulation run. The simulated systems valid-
ity was checked by evaluating the quality check parameters
(temperature, pressure, kinetic energy) during the MD run. It was
observed that stability was maintained in each system.
After periodic boundary correction of each simulated tra-
jectory, the stability of the systems was measured by calcu-
lating backbone RMSD (root mean square deviation) during
Table 1. List of phyto-compounds with docking score (CTX-M-3, CTX-M-15) and ADMET properties.
Compounds
Docking Score (kcal/mol)
G.I.Absorption (%) BBB
No of
violation AMES
Hepato
toxicity
LD50
(mol/kg)
Drug
likeness SA PAINSCTX-M-3 CTX-M-15
Catechin gallate (CG) 8.0 8.2 62 No 1 No No 2.55 0.93 4.16 One alert
Luteolin (Lut) 7.9 7.8 81 No No No No 2.45 0.38 3.02 One alert
Uvaol (Uva) 7.0 7.1 92.8 No 1 No No 2.64 0.2 6.28 No
Silibinin (Sil) 8.8 9.0 62 No No No No 2.55 0.84 4.92 No
Cefotaxime (CTX) 6.2 6 38 No 1 No Yes 1.93 1.37 4.79 No
JOURNAL OF BIOMOLECULAR STRUCTURE AND DYNAMICS 7

the 200 ns MD run. The radius of gyrate (Rg) indicates the
structural compactness of the system. Any deviation signifies
unstable folding, and uniformity indicates stable structure
folding during an MD run (Rhodes, 2006).
While calculating the backbone RMSD of the apo-proteins
and complexes, the lowest RMSD value indicates higher stability.
The average backbone, RMSD of CTX-M-3 and CTX-M-15 apo-
protein were 0.06 nm and 0.084 nm, respectively (Figure 2A). For
uvaol (Uva) highest fluctuation was observed for both proteins.
For the last 80 ns, the average RMSD values with CTX-M-3 and
CTX-M-15 protein were 0.17 nm and 0.19 nm, respectively. For
CG, silibinin, and luteolin, a similar trend of RMSD was observed
with an average of 0.13 nm. For cefotaxime (CTX), a steep peak
was observed near 120 ns with CTX-M-15 protein. From starting
to 100 ns, the average RMSD value was 0.108 nm, but from 100
to 200 ns, the average rose to 0.25 nm (Figure 2B). The stability
in ligand RMSD values indicated that all the phyto-compounds
were maintained in the binding site throughout the 200 ns MD
run (Figure Supplementary S5).
While calculating Rg values, no significant fluctuation was
observed with CTX-M-3 protein, except for uvaol. The average Rg
value for complexes with CTX-M-3 was 1.82 nm, and silibinin had
the lowest average Rg(1.76 nm). For CTX-M-3-Uvaol, a sudden
shift was observed around 170ns, but after that, it achieved
stability (Figure 2C). For CTX-M-15 protein, the lowest average Rg
value was with silibinin (1.68nm), but the rest had an average
Rg value of 1.82 nm (Figure 2D). The lack of significant fluctu-
ation in the Rg graph reflects no protein instability while folding
during the MD run. As the Rg value depicts the root mean
square average of the distance of the scattering atoms from the
mass center of the molecule, less Rg value of silibinin with both
proteins reflects less scattering of atoms and more stability of
the complex structure.
3.5. Analysis of residual fluctuations (RMSF) and solvent
accessible of surface area (SASA)
After calculating structures’RMSD and Rg values, we calcu-
lated root mean square fluctuation (RMSF), which denotes
residual fluctuation during the MD run. Higher fluctuation
indicates higher structural mobility. For CTX-M-3 protein,
high peaks were observed for both luteolin and uvaol. CG
and silibinin appeared to be stable (Figure 3A). While observ-
ing RMSF values for the active site residues, we can observe
that for uvaol highest fluctuation was observed in Tyr80 and
Thr191 residues (Supplementary Figure S3A). For the CTX-M-
15 protein, the fluctuation was observed in 80-120 and 210-
230 residues for all the complexes (Figure 3B). Here also,
Figure 2. Dynamics stabilities of complex and apo-protein systems during 200 ns MD simulation. (A) Backbone RMSD of systems compared to the initial structure
during 200 ns MD (B) The radius of gyration of complex systems and the apo-protein.
8 B. KAR ET AL.

uvaol appeared to have the highest flexibility among all the
residues during the MD run. But in the active site, uvaol did
not show much flexibility, which perfectly corroborates the
previous result of loose binding with the target protein
(Supplementary Figure S3B).
For calculating solvent accessible surface area (SASA), the
higher value represents more surface area of the molecule
exposed to solvent. The lower value represents the ligand is
deeply buried inside the active site pocket of the protein. The
average SASA values of CTX-M-3 and CTX-M-15 proteins were
113.9nm
2
and 121.4nm
2,
respectively. The average SASA value
for CTX-M-3-Silibinin was 118.3nm
2
; for uvaol, a rise was
observed near 150 ns (average for last 50 ns: 120.8 nm
2
)(Figure
3C). The average SASA value for CTX-M-15 in complex with uvaol
and CTX were 121.7 nm
2
and 128.6 nm
2
(Figure 3D). So about
the previous results, SASA graphs confirmed that, among phyto-
compounds, uvaol is not deeply buried inside the target protein
and therefore had the lowest binding affinity with the target
protein.
3.6. Hydrogen bond dynamics
Intermolecular H bond analysis plays a vital role in determin-
ing complex stability. The average H bond of CG with both
CTX-M proteins was 4, but with CTX-M-15 protein, till 100 ns,
the average H bond was 6, but after that, some distortion
could be observed. Continuity could be observed throughout
the MD run for both CTX-M-3 and CTX-M-15 proteins with
silibinin (Figure 4). Due to this continuity, the stronger bind-
ing affinity between silibinin and target proteins could be
explained. The average lowest H bond throughout the 200 ns
MD run could be observed for uvaol.
While understanding the dynamics of H bonds during the
MD run from the graph, we could observe distortions that hydro-
phobic and electrostatic interactions could further replace. But
the total H bond number and the residues occupied during H
bonding could not be inferred from the graph. So we tried to
calculate H bond occupancy for each complex. For CTX-M-3-CG,
the total H bond was 50, and the highest occupancy was
observed with Asp250 (33.83%). Reference drug cefotaxime (CTX)
hadformedthemaximumretainedHbond(41.41%)with
Glu141. Although the total number of H bonds (58) was high
with silibinin, the highest occupied H bond was observed
with Trp185 with 4.59% (Table 2).
For CTX-M-15 protein, the highly occupied H bond was
observed with CG (91.46%). For CTX-M-15-Silibinin, the total
number of H bonds was 62, and the high occupied H bond
(12.43%) was with Asp131. The lowest binding affinity could be
justified with the lowest number of H bonds formed between
uvaol and CTX-M-15 protein. H bond dynamics and occupancy
data perfectly corroborate the previous MD trajectory analysis.
3.7. Principal component analysis (PCA)
Principal component analysis of the MD trajectories was per-
formed to understand the collective motion of the complex
Figure 3. (A) RMSF analysis of CTX-M-3 apo-protein and complexes. (B) RMSF analysis of CTX-M-15 in complex with phyto-compounds and CTX. (C) Time-wise
SASA calculation during 200 ns MD run of CTX-M-3 complexes. (D) Time-wise SASA calculation during 200 ns MD run of CTX-M-15 complexes.
JOURNAL OF BIOMOLECULAR STRUCTURE AND DYNAMICS 9

Figure 4. Dynamics of inter-molecular H-bonds complex/Apo-protein systems over 200 ns MD run. (A) Dynamics of H-bonds for CTX-M-3 complexes. (B) H bond
dynamics of CTX-M-15 complexes.
Figure 5. (A, C) PCA scatter plot representation. (A) Scatter plot representation of CTX-M-3 protein and complexes. (B) Scatter plot representation of CTX-M-15 pro-
tein and complexes. (B, D) Eigenvalues with respect to eigenvector indices of the six systems investigated in the present study. (B) Graphical Plot of first two twenty
eigenvalues of CTX-M-3 protein and complexes. (D) Graphical Plot of the first twenty eigenvectors of CTX-M-15 and complexes.
10 B. KAR ET AL.

structures and compare them with the apo-protein. It was
performed by constructing a diagonal covariance matrix by
selecting the Caatom of the structures. The covariance
matrix captures the global motion of the atoms through
principal components/eigenvectors. Each eigenvector repre-
sent a specific eigenvalue. Eigenvectors describe the atoms’
global direction of motion, and eigenvalues represent the
motion contribution by atoms in the MD trajectory. The first
few eigenvectors represent a large amount of internal
motion of the atoms in each system. The first twenty eigen-
values were plotted for each system better to understand
conformational changes in the degree of fluctuation.
Subsequently, a scatter plot is generated by projecting the
first two eigenvectors (PC) into the phase space. Higher dis-
persion in the scatter plot represents more degree of
flexibility.
For CTX-M-3 apo-protein and complexes, the highest dis-
persion was observed with uvaol, followed by cefotaxime
(CTX). The lowest dispersion was observed for CG and silibi-
nin, which means both systems had achieved stability during
the MD run (Figure 5A). The high dispersion of uvaol and
cefotaxime are in correspondence with the previous RMSD
plot, which entails structural instability. While plotting the
first twenty eigenvectors, the highest peak was observed
with luteolin, followed by cefotaxime and uvaol. But after
the first eigenvalue, a sudden decline could be observed,
which can be concluded that the complex luteolin system
was unstable during the first few ns of the MD run. The
RMSF plot also suggested a few high fluctuated peaks in
CTX-M-3-Luteolin complex system (Figure 5B).
Among the different complexes, CTX-M-15-CTX and CTX-M-
15-Uvaol represented the highest dispersion, and CTX-M-15-
CG displayed the lowest dispersion (Figure 5C). The trace
value of covariance matrices was found as CTX-M-15-CTX
(17.34nm
2
), followed by CTX-M-15-Uvaol (15.90nm
2
), CTX-M-
15-Luteolin (10.27nm
2
), CTX-M-15 (9.53nm
2
), CTX-M-15-
Silibinin (8.68nm
2
), and CTX-M-15-CG (6.90nm
2
). The scatter
plot representation of CTX-M-15 apo-protein and complexes
are totally in agreement with the previous trajectory data ana-
lysis. CTX-M-15-Uvaol had the highest dispersion, which is in
agreement of previous RMSD plot, signifies instability.
A sharp decline was observed after the first eigenvalue in
CTX M-15-Silibinin, CTX-M-15-CTX, and CTX-M-15-Uvaol to
reach some constrained and more localized fluctuation
(Figure 5D). For CTX-M-15-Silibinin, after the first eigenvector,
the rest were seen to achieve stability. After the initial
instability, the system achieved stability. The flexibility of the
active site residues confirmed their binding As in CTX-M-15-
Silibinin, the first twenty eigenvectors occupy 91.3% (the
highest), followed by CTX-M-15-Uvaol (85.3%) and CTX-M-15-
Luteolin (85%). For the apo-protein (CTX-M-15), the first
twenty eigenvectors occupy 80.6% of the total backbone
motion.
3.8. Dynamic cross-correlation map (DCCM)
The effect of correlated interaction between the residues of
the systems during the MD run can be analyzed through
DCCM analysis. The matrix represented the correlated and
anti-correlated motions by cyan and pink, respectively
(Figure 6). For CTX-M-3 apo-protein, the matrix does not
show significantly correlated motions between the amino
acid residues, but in CTX-M-3-CG, increased anti-correlation
could be observed between 120-160 residues. In CTX-M-3-
Silibinin, the anti-correlation is not dominant, but a distinct
correlation could be observed near 50-120 residues. If we
check the active site residues, Asn79, Tyr80, Ile83, Tyr104,
and Asn105 lie in the same region, and due to silibinin bind-
ing, the correlated motion has increased in those regions.
For CTX-M-15 protein, as compared to CTX-M-3, an increased
anti-correlation (pink) could be observed, but due to CG
binding, the anti-correlation between the amino acid resi-
dues has fainted. Due to silibinin binding, an increased corre-
lated motion we could observe. Especially from 150-170 and
200-270 residues, a distinct correlation (cyan) in amino acid
residues is confirmed. If we check the protein structure,
Table 2. H bond occupancy of the complex structures.
Complex Donor Acceptor Occupancy Complex Donor Acceptor Occupancy
CTX-M-3-CG (50 H-bonds) UNK0 Asp250 33.83% CTX-M-15-Catechin Gallate (54 bonds) UNK0 Glu166 91.46%
UNK0 Glu11 32.14% UNK0 Ala270 10.78%
UNK0 Asp214 13.27% Ser237 UNK0 6.49%
UNK0 Glu141 8.14% UNK0 Thr216 4.12%
CTX-M-3-CTX (34 H-bonds) UNK0 Glu141 42.42% CTX-M-15-Cefotaxime (80 bonds) Val231 UNK0 5.50%
Thr191 UNK0 10.62% UNK0 Trp229 2.11%
UNK0 Asp214 4.34% Thr215 UNK0 2.06%
Ser105 UNK0 2.70% Thr216 UNK0 1.99%
CTX-M-3-Luteolin (40 H-bonds) UNK0 Glu11 5.93% CTX-M-15-Luteolin (65 H-bonds) UNK0 Ser237 3.69%
UNK0 Glu62 5.12% UNK0 Glu166 3.40%
UNK0 Asp214 1.56% UNK0 Glu110 2.80%
UNK0 Glu13 1.48% UNK0 Glu271 1.55%
CTX-M-3-Silibinin (58 H bonds) Trp185 UNK0 4.59% CTX-M-15-Silibinin (62 H bonds) UNK0 Asp131 12.43%
His87 UNK0 2.89% UNK0 Glu121 1.34%
UNK0 Glu96 2.59% Val231 UNK0 0.72%
Glu103 UNK0 1.21% UNK0 Lys111 0.67%
CTX-M-3-Uvaol (26 H bonds) Ser45 UNK0 12.89% CTX-M-15-Uvaol (33 H bonds) UNK0 Asp163 3.69%
UNK0 Asn107 12.47% UNK0 Val103 3.44%
Ser105 UNK0 3.88% UNK0 Ser237 0.54%
UNK0 Asp106 1.35% UNK0 Lys129 0.54%
JOURNAL OF BIOMOLECULAR STRUCTURE AND DYNAMICS 11

these amino acids reside in the helical loop, actively partici-
pating in ligand binding. The rest of the systems did not
have significant alteration than apo-proteins.
3.9. Binding free energy analysis
In drug discovery, virtual screening of inhibitors through
molecular docking and molecular dynamics-based simulation
is crucial to identify the potential leads. These computational
approaches help researchers identify the best docking ligand
pose, protein conformational changes due to ligand binding,
etc. But each complex’s energetic contribution to the resi-
dues may narrate a different story. Thus binding free energy
of each protein-ligand complex system using 1000 snapshots
of the last 100 ns can be calculated using the end-point
approach MM/PBSA. The binding free energy of each com-
plex is depicted in (Table 3).
The binding free energy of CTX-M-3-Silibinin (-191.557þ/-
4.596 kJ/mol) is higher than CTX-M-3-CG (-140.263þ/-
1.104 kJ/mol). As justified by the previous docking and MD
trajectory analysis, silibinin had a higher binding affinity with
CTX-M-3 protein. Due to binding with the protein, the stabil-
ity and correlated motions in the molecule increased. The
binding free energy of CTX-M-15-Silibinin (-203.557þ/-
4.596 kJ/mol) and CTX-M-15-Luteolin (-162.177þ/-
1.209 kJ/mol) were higher in comparison with others. It was
observed that, in the calculation of binding free energy, van
der Waals energy was the most contributing factor in each
complex. Polar solvation energy had a positive value, oppos-
ing the other two (electrostatic, SASA) favorable contributing
factors for final, binding free energy. Next, analysis was per-
formed to understand the energetic contribution of the
active site residues in ligand binding per residue decompos-
ition. In the case of CTX-M-15-Silibinin, six H bonds (Ser70,
Asn132, Ser237, Gly241, Thr242, Pro268), ten hydrophobic
interactions (Asn104, Ser130, Asn170, Thr216, Thr235, Ser237,
Gly238, Gly239, Lys269, Ala270) and one (Tyr105) pi-pi con-
tact was observed. In the case of CTX-M-15-CTX, two residues
Figure 6. Dynamic cross-correlation maps (DCCM) of (A) CTX-M-3 (Apo-protein), (B) CTX-M-3-CG, (C) CTX-M-3-Sil, (D) CTX-M-15, (E) CTX-M-15-CG, and (F) CTX-M-
15-Sil.
12 B. KAR ET AL.

(Lys73 and Lys234) were involved in salt bridge formation.
We can conclude from both CTX-M-3 and CTX-M-15 protein
complexes that the active site residues and the adjoining
helical loop regions actively participate in ligand binding
(Supplementary Figure S4).
3.10. Confirmation of CTX-M-3 and CTX-M-15 producing
E.coli bacterial strains
Bacterial clinical isolates were identified through a (Vitek2
compact) automatic bacterial identification system. Antibiotic
susceptibility tests were performed using discs of different
classes of antibiotics like aminoglycosides (amikacin, genta-
micin), beta-lactams(ampicillin, cephalosporins, carbape-
nems), fluoroquinolones (ciprofloxacin, ofloxacin,
levofloxacin). From them, isolates that were resistant to
cephalosporins but susceptible to carbapenems (imipenem)
were selected. A double-disc diffusion test was for ESBL
activity. The increased zone of inhibition was measured for
antibiotics in combination with an inhibitor (clavulanic acid)
compared to antibiotics alone (cefotaxime). Subsequently, to
confirm the ESBL genes selected isolates were screened
through polymerase chain reaction (PCR). Three CTX-M-3
positive bacterial isolates (AMCH92, AMCH142, UTI2) and
three CTX-M-15 producing (AMCH1, AMCH33, AMCH68) were
selected for the in vitro antibacterial activity testing.
3.11. Calculation of minimum inhibitory concentration
(MIC) and minimum bactericidal
concentration (MBC)
After the virtual screening, the same four phyto-compounds
were selected to check their antibacterial activity against
bacterial strains. Six bacterial strains used for the experiment
were highly resistant to antibiotics (cefotaxime) and suscep-
tible to imipenem (no growth was observed till 0.25 mg/mL).
The phyto-compounds were tested at 1024-2mg/mL to check
their antibacterial activity. As expected from the in silico study,
uvaol showed no significant antibacterial activity. For catechin
gallate and luteolin, MIC was 512 mg/mL; for silibinin, the MIC
was 128 mg/mL. MBC of silibinin in each isolate was 256 mg/mL
(Table 4). MBC of catechin gallate and luteolin were 4096mg/mL
and 2048 mg/mL, respectively. The ratio between MBC/MIC if 
4, and the compound is considered bactericidal (Mogana et al.,
2020). Thus from the in vitro antibacterial assay, we can conclude
that both luteolin and silibinin have a bactericidal effect. Still,
considering the MIC, silibinin has better antibacterial activity than
luteolin. Literature reviews confirmed catechin gallate as a PAINS
compound, thus showing good in silico results, but the com-
pound failed to provide satisfactory results in vitro (J. B. Baell,
2016).
3.12. Calculation of fractional inhibitory
concentration (RFIC)
After checking the individual antibacterial activity of the phyto-
compounds, silibinin, and luteolin were tested for the antibac-
terial activity of the compounds in combination with antibiot-
ics. The combination assay revealed the synergistic activity of
the phyto-compounds by calculating FIC. If the phyto-com-
pound appears to be synergistic, then the phyto-compound
can improve the market shelf life of antibiotics after being
administered in combination. Here the same six isolates were
chosen for the combination assay. For all six isolates, the syner-
gistic activity of the silibinin is clear with CTX, but for luteolin,
only one isolate (AMCH92) showed synergistic activity (Table
5). Thus we can confirm that silibinin alone and in combination
with cefotaxime have good anti-ESBL training.
3.13. Nitrocefin assay
Nitrocefin assay was conducted to check the direct beta-lac-
tamase activity of silibinin. As expected, gentamicin did not
show any inhibitory activity for beta-lactamase, but clavu-
lanic acid is reported to have an inhibitory effect in both liv-
ing cells and dead bacterial lysate. Silibinin at MIC/2
concentration showed beta-lactamase inhibitory activity like
clavulanic acid, but only in living bacterial cells. Silibinin is
not effective against dead bacterial cells. Thus, we can con-
clude that silibinin requires living cells to prevent enzymes
and does not have direct beta-lactamase inhibitory activity
like clavulanic acid. The exact mechanism needs to be
explored more (Figure 7).
Thus from the present study, we can conclude that silibi-
nin has anti-ESBL activity, but it needs further investigation
for confirmation. Silibinin phyto-compound is derived from
the milk thistle plant and is known for its anti-inflammatory
activity by inhibiting the TLR4 protein (Song et al., 2016). It is
approved in Europe as an intravenous drug (Legalon-SIL) for
treating hepatic failure caused by amatoxin-releasing
Table 3. The Binding free energy of each complex is calculated using MM/PBSA method.
Complex
Van der Waal energy
(kJ/mol)
Electrostatic energy
(kJ/mol)
Polar solvation energy
(kJ/mol)
SASA energy
(kJ/mol)
Binding free energy
(kJ/mol)
CTX-M-3-CG 112.832þ/- 2.499 94.63þ/1.867 72.826þ/- 2.777 5.627þ/-1.500 140.263þ/- 1.104
CTX-M-3-Luteolin 109.718þ/- 1.764 75.78þ/- 2.190 96.812þ/- 2.650 15.47þ/-1.512 104.158þ/-1.209
CTX-M-3-Silibinin 107.519þ/- 1.780 147.190þ/-3.015 73.188þ/- 0.279 10.036þ/0.028 191.557þ/- 4.596
CTX-M-3-Uvaol 86.95þ/- 3.154 70.25þ/- 2.101 82.50þ/- 0.452 6.07þ/- 0.490 80.70þ/- 4.487
CTX-M-3-CTX 156.982þ/- 2.632 102.732þ/-1.367 179.929þ/-1.394 3.486þ/- 0.395 83.271þ/-3.202
CTX-M-15-CG 206.931þ/- 2.499 205.382þ/1.867 283.418þ/- 2.777 14.748þ/-1.500 143.643þ/- 1.104
CTX-M-15-Luteolin 190.718þ/- 1.764 157.79þ/- 2.190 196.81þ/- 2.650 10.482þ/-1.512 162.177þ/-1.209
CTX-M-15-Silibinin 179.519þ/- 1.780 177.190þ/-3.015 173.188þ/- 0.279 20.036þ/0.028 203.557þ/- 4.596
CTX-M-15-Uvaol 46.956þ/- 3.154 13.773þ/- 0.101 29.717þ/- 0.452 5.997þ/- 0.490 27.009þ/- 4.487
CTX-M-15-CTX 166.982þ/- 2.632 111.734þ/-1.367 142.909þ/-1.394 5.593þ/- 0.395 141.398þ/-3.202
JOURNAL OF BIOMOLECULAR STRUCTURE AND DYNAMICS 13

mushroom ingestion. In emergencies, it can be administrated
to patients in the U.S. with FDA approval. Clinical trials
proved no significant toxicity or side effects were associated
with Silibinin (Mengs et al., 2012). Thus repurposing silibinin
can be an excellent therapeutic option for treating patients
with CTX-M-3 and CTX-M-15-producing bacterial infections.
4. Conclusion
In the present day, virtual screening is a practical approach to
identifying potential leads in drug discovery. A combined strategy
of bioinformatics, medicinal chemistry, and biotechnology is
essential to fight the pace of emerging antibacterial resistance. In
the present study, we have developed a virtual screening work-
flow to identify potential inhibitors of CTX-M-3 and CTX-M-15 pro-
teins. Following molecular docking and ADMET property
screening, an MD simulation was performed of the selected four
compounds and compared them with cefotaxime and apo-pro-
tein. After binding free energy calculation, it was evident that sili-
binin was the best potential lead, among others. After in silico
analysis, in vitro antibacterial validated the above result. Therefore
the present study proclaims the ESBL inhibitory potential of silibi-
nin, and our study is the first time reporting silibinin as both CTX-
M-3 and CTX-M-15 inhibitor. But before moving further into the
clinical trial, we should verify the potency of silibinin. Still, the pre-
sent study confirmed that silibinin could be treated as a potential
lead in drug designing against CTX-M enzymes. The adapted
workflow will help others to screen potential compounds and
develop drugs to fight against antibiotic resistance.
Acknowledgments
The authors gratefully acknowledge all the healthcare workers for their tireless
dedication at each level to fight COVID-19. The authors thank the Indian
Council of Medical Research, New Delhi, for providing financial support for the
study through intramural funding and the Council of Scientific and Industrial
Research for providing a senior fellowship to the 1st author. The authors also
acknowledgethesupportbytheDepartmentofHealthResearch,Ministryof
Health and Family Welfare, Govt. of India for providing financial support to BD
through the DHR-Young Scientist Grant.
Disclosure statement
No potential conflict of interest was reported by the authors.
Ethics approval
The study was ethically approved by the institutional human ethical
committee of ICMR—Regional Medical Research Centre, Bhubaneswar.
Table 4. MIC and MBC of phyto-compounds and antibiotics for six CTX-M-producing bacterial isolates.
Phyto-compounds
(mg/mL)
MBC / MIC
CTX-M-15 CTX-M-3
AMCH 1 AMCH 33 AMCH 68 AMCH 92 AMCH 142 UTI 2
Catechin Gallate MIC 512 MIC 512 MIC 512 MIC 512 MIC 512 MIC 512 8
MBC 4096 MBC 4096 MBC 4096 MBC 4096 MBC 4096 MBC 4096
Luteolin MIC 512 MIC 512 MIC 512 MIC 512 MIC 512 MIC 512 4
MBC 2048 MBC 2048 MBC 2048 MBC 2048 MBC 2048 MBC 2048
Silibinin MIC 128 MIC 128 MIC 128 MIC 128 MIC 128 MIC 128 2
MBC 256 MBC 256 MBC 256 MBC 256 MBC 256 MBC 256
Uvaol MIC >1024 MIC >1024 MIC >1024 MIC >1024 MIC >1024 MIC >1024 Not Found
Table 5. Analysis of FIC of luteolin and silibinin in combination with cefotaxime (CTX).
Bacterial Strains
MIC (mg/mL)
Index (Interpretation)
MIC (mg/mL)
Index (Interpretation)CTX Luteolin CTX Silibinin
AMCH 1 1024 512 0.75 (Additive) 512 128 0.37 (Synergy)
AMCH 33 2048 512 0.625 (Additive) 1024 128 0.25 (Synergy)
AMCH 68 1024 512 0.75 (Additive) 512 128 0.37 (Synergy)
AMCH 92 1024 512 0.5 (Synergy) 512 128 0.37 (Synergy)
AMCH 142 1024 512 0.625 (Additive) 512 128 0.37 (Synergy)
UTI2 2048 512 0.75 (Additive) 1024 128 0.31 (Synergy)
Figure 7. Beta-lactamase inhibitory potential of silibinin in both living bacterial cells and bacterial lysates.
14 B. KAR ET AL.

Funding
This work was supported by CSIR-SRF.
ORCID
Debdutta Bhattacharya http://orcid.org/0000-0001-5199-5288
References
Anand, U., Jacobo-Herrera, N., Altemimi, A., & Lakhssassi, N. (2019). A
comprehensive review on medicinal plants as antimicrobial therapeu-
tics : Potential avenues of biocompatible drug discovery. Metabolites,
9(11), 258–213. https://doi.org/10.3390/metabo9110258
Arnold, K., Bordoli, L., Kopp, J., & Schwede, T. (2006). The SWISS-MODEL
workspace: A web-based environment for protein structure homology
modelling. Bioinformatics,22(2), 195–201. https://doi.org/10.1093/bio-
informatics/bti770
Baell, J. B. (2016). Feeling nature’s PAINS: Natural products, natural prod-
uct drugs, and pan assay interference compounds (PAINS). Journal of
Natural Products,79(3), 616–628. https://doi.org/10.1021/acs.jnatprod.
5b00947
Baell, J., & Walters, M. A. (2014). Chemical con artists foil drug discovery.
Nature,513(7519), 481–483. https://doi.org/10.1038/513481a
Benkert, P., Biasini, M., & Schwede, T. (2011). Toward the estimation of
the absolute quality of individual protein structure models.
Bioinformatics,27(3), 343–350. https://doi.org/10.1093/bioinformatics/
btq662
Berendsen, H. J. C., Postma, J. P. M., Van Gunsteren, W. F., Dinola, A., &
Haak, J. R. (1984). Molecular dynamics with coupling to an external
bath. The Journal of Chemical Physics,81(8), 3684–3690. https://doi.
org/10.1063/1.448118
Berendsen, H. J. C., van der Spoel, D., & van Drunen, R. (1995).
GROMACS: A message-passing parallel molecular dynamics implemen-
tation. Computer Physics Communications,91(1–3), 43–56. https://doi.
org/10.1016/0010-4655(95)00042-E
Bevan, E. R., Jones, A. M., & Hawkey, P. M. (2017). Global epidemiology
of CTX-M b-lactamases: Temporal and geographical shifts in geno-
type. The Journal of Antimicrobial Chemotherapy,72(8), 2145–2155.
https://doi.org/10.1093/jac/dkx146
Camacho, C., Coulouris, G., Avagyan, V., Ma, N., Papadopoulos, J., Bealer,
K., & Madden, T. L. (2009). BLASTþ: Architecture and applications. BMC
Bioinformatics,10(1), 1–9. https://doi.org/10.1186/1471-2105-10-421
Cant
on, R., Gonz
alez-Alba, J. M., & Gal
an, J. C. (2012)). CTX-M enzymes:
Origin and diffusion. Frontiers in Microbiology,3, 1-21. https://doi.org/
10.3389/fmicb.2012.00110
Castanheira, M., Simner, P. J., & Bradford, P. A. (2021). Extended-spectrum
b-lactamases: An update on their characteristics, epidemiology and
detection. JAC-Antimicrobial Resistance,3(3), 1-21. https://doi.org/10.
1093/jacamr/dlab092
Chen, J., Zeng, Q., Wang, W., Sun, H., & Hu, G. (2022). Decoding the iden-
tification mechanism of an SAM-III riboswitch on ligands through mul-
tiple independent Gaussian-accelerated molecular dynamics
simulations. Journal of Chemical Information and Modeling,62(23),
6118–6132. https://doi.org/10.1021/acs.jcim.2c00961
Chen, J., Zhang, S., Wang, W., Pang, L., Zhang, Q., & Liu, X. (2021).
Mutation-induced impacts on the switch transformations of the GDP-
and GTP-bound K-ras: Insights from multiple replica Gaussian acceler-
ated molecular dynamics and free energy analysis. Journal of
Chemical Information and Modeling,61(4), 1954–1969. https://doi.org/
10.1021/acs.jcim.0c01470
Chen,V.B.,Arendall,W.B.,Headd,J.J.,Keedy,D.A.,Immormino,R.M.,
Kapral,G.J.,Murray,L.W.,Richardson,J.S.,&Richardson,D.C.(2010).
MolProbity: All-atom structure validation for macromolecular crystallog-
raphy. Acta Crystallographica Section D Biological Crystallography,66(1),
12–21. https://doi.org/10.1107/S0907444909042073
Colovos, C., & Yeates, T. O. (1993). Verification of protein structures:
Patterns of nonbonded atomic interactions. Protein Science,2(9),
1511–1519. https://doi.org/10.1002/pro.5560020916
Daina, A., Michielin, O., & Zoete, V. (2017). SwissADME: A free web tool
to evaluate pharmacokinetics, drug-likeness and medicinal chemistry
friendliness of small molecules. Scientific Reports,7(1), 1–13. https://
doi.org/10.1038/srep42717
Dehury, B., Patra, M. C., Maharana, J., Sahu, J., Sen, P., Modi, M. K.,
Choudhury, M. D., & Barooah, M. (2014). Structure-based computa-
tional study of two disease resistance gene homologues (Hm1 and
Hm2) in maize (Zea mays L.) with implications in plant-pathogen.
PLoS One,9(5), e97852. https://doi.org/10.1371/journal.pone.0097852
Dhillon, R. H. P., & Clark, J. (2012). ESBLs: A clear and present danger?
Critical Care Research and Practice,2012, 625170. https://doi.org/10.
1155/2012/625170
Doak, B. C., Giordanetto, F., & Kihlberg, J. (2014). Review oral druggable
space beyond the rule of 5: Insights from drugs and clinical candi-
dates. Chemistry & Biology, 1115–1142. https://doi.org/10.1016/j.chem-
biol.2014.08.013
Doi, Y., Iovleva, A., & Bonomo, R. A. (2017). The ecology of extended-
spectrum b-lactamases (ESBLs) in the developed world. Journal of
Travel Medicine,24(suppl_1), S44–S51. https://doi.org/10.1093/jtm/
taw102
Dutour, C., Bonnet, R., Marchandin, H., Boyer, M., Chanal, C., Sirot, D., &
Sirot, J. (2002). CTX-M-1, CTX-M-3, and CTX-M-14 -lactamases from
enterobacteriaceae isolated in France. Society,46(2), 534–537. https://
doi.org/10.1128/AAC.46.2.534
EClinicalMedicine. (2021). Antimicrobial resistance: A top ten global pub-
lic health threat. EClinicalMedicine,41, 101221. https://doi.org/10.1016/
j.eclinm.2021.101221
Ertl, P., & Schuffenhauer, A. (2009). Estimation of synthetic accessibility
score of drug-like molecules based on molecular complexity and frag-
ment contributions. Journal of Cheminformatics,1(1), 1–11. https://doi.
org/10.1186/1758-2946-1-8
Ewald, P. P. (1921). Die Berechnung optischer und elektrostatischer
Gitterpotentiale. Annalen Der Physik,369(3), 253–287. https://doi.org/
10.1002/andp.19213690304
Fair, R. J., & Tor, Y. (2014). Antibiotics and bacterial resistance in the 21st
century. Perspectives in Medicinal Chemistry,6, PMC.S14459–64.
https://doi.org/10.4137/PMC.S14459
Ferreira De Freitas, R., & Schapira, M. (2017). A systematic analysis of
atomic protein-ligand interactions in the PDB. MedChemComm,8(10),
1970–1981. https://doi.org/10.1039/C7MD00381A
Govindaswamy, A., Bajpai, V., Khurana, S., Batra, P., Mathur, P., &
Malhotra, R. (2019). Prevalence and characterization of beta-lacta-
mase-producing Escherichia coli isolates from a tertiary care hospital
in India. Journal of Laboratory Physicians,11(3), 123–127. https://doi.
org/10.4103/jgid.jgid_68_18
Grant, B. J., Rodrigues, A. P. C., ElSawy, K. M., McCammon, J. A., & Caves,
L. S. D. (2006). Bio3d: An R package for the comparative analysis of
protein structures. Bioinformatics,22(21), 2695–2696. https://doi.org/
10.1093/bioinformatics/btl461
Guan, L., Yang, H., Cai, Y., Sun, L., Di, P., Li, W., Liu, G., & Tang, Y. (2019).
ADMET-score-a comprehensive scoring function for evaluation of
chemical drug-likeness. MedChemComm,10(1), 148–157. https://doi.
org/10.1039/C8MD00472B
Hanwell, M. D., Curtis, D. E., Lonie, D. C., Vandermeersch, T., Zurek, E., &
Hutchison, G. R. (2012). Avogadro: An advanced semantic chemical
editor, visualization, and analysis platform. Journal of Cheminformatics,
4(17), 1–17. https://doi.org/10.1016/j.aim.2014.05.019
Hess, B., Bekker, H., Berendsen, H. J. C., & Fraaije, J. G. E. M. (1997).
LINCS: A linear constraint solver for molecular simulations. Journal of
Computational Chemistry,18(12), 1463–1472. https://doi.org/10.1002/
(SICI)1096-987X(199709)18:12<1463::AID-JCC4>3.0.CO;2-H
Huang, J., & MacKerell, A. D., Jr. (2013). CHARMM36 all-atom additive
protein force field: Validation based on comparison to NMR data.
Journal of Computational Chemistry,34(25), 2135–2145. https://doi.
org/10.1002/jcc.23354
JOURNAL OF BIOMOLECULAR STRUCTURE AND DYNAMICS 15

Hutchings, M., Truman, A., & Wilkinson, B. (2019). Antibiotics: Past, pre-
sent and future. Current Opinion in Microbiology,51(1), 72–80. https://
doi.org/10.1016/j.mib.2019.10.008
K., Vanommeslaeghe, E., Hatcher, C., Acharya, S., Kundu, S., Zhong, J.,
Shim, E. D. O., Guvench, P., Lopes, I., & Vorobyov, A. D. M. J. (2012).
CHARMM general force field: A force field for drug-like molecules
compatible with the CHARMM all-atom additive biological force fields
K. Journal of Computational Chemistry,32, 174–182. https://doi.org/10.
1002/jcc
Kadri, S. S. (2020). Key takeaways from the U.S. CDC’s 2019 antibiotic
resistance threats report for frontline providers. Critical Care Medicine,
2, 939–945. https://doi.org/10.1097/CCM.0000000000004371
Kar, B., Sharma, M., Peter, A., Chetia, P., Neog, B., Borah, A., Pati, S., &
Bhattacharya, D. (2021). Prevalence and molecular characterization of
b-lactamase producers and fluoroquinolone resistant clinical isolates
from North East India. Journal of Infection and Public Health,14(5),
628–637. https://doi.org/10.1016/j.jiph.2021.02.007
Khameneh, B., Iranshahy, M., Soheili, V., Sedigheh, B., & Bazzaz, F. (2019).
Review on plant antimicrobials: A mechanistic viewpoint.
Antimicrobial Resistance & Infection Control,8.
Kumari, R., Kumar, R., & Lynn, A. (2014). G-mmpbsa -A GROMACS tool for
high-throughput MM-PBSA calculations. Journal of Chemical
Information and Modeling,54(7), 1951–1962. https://doi.org/10.1021/
ci500020m
Lahiri, S. D., Mangani, S., Durand-Reville, T., Benvenuti, M., De Luca, F.,
Sanyal, G., & Docquier, J. D. (2013). Structural insight into potent
broad-spectrum inhibition with reversible recyclization mechanism:
Avibactam in complex with CTX-M-15 and pseudomonas aeruginosa
AmpC b-lactamases. Antimicrobial Agents and Chemotherapy,57(6),
2496–2505. https://doi.org/10.1128/AAC.02247-12
Lai, C.-C., Chen, S.-Y., Ko, W.-C., & Hsueh, P.-R. (2021). Combating anti-
microbial resistance during the COVID-19 pandemic. International
Journal of Antimicrobial Agents,57(4), 106324–106398. https://doi.org/
10.1016/j.ijantimicag.2021.106324
Laskowski, R. A., & Swindells, M. B. (2011). LigPlotþ: Multiple ligand-pro-
tein interaction diagrams for drug discovery. Journal of Chemical
Information and Modeling,51, 2778–2786.
Laskowski, R. A., MacArthur, M. W., Moss, D. S., & Thornton, J. M. (1993).
PROCHECK: A program to check the stereochemical quality of protein
structures. Journal of Applied Crystallography,26(2), 283–291. https://
doi.org/10.1107/S0021889892009944
Lipinski, C. A., Lombardo, F., Dominy, B. W., & Feeney, P. J. (2012).
Experimental and computational approaches to estimate solubility
and permeability in drug discovery and development settings.
Advanced Drug Delivery Reviews,64,4–17. https://doi.org/10.1016/j.
addr.2012.09.019
Liu, S., Zhou, Y., Niu, X., Wang, T., Li, J., Liu, Z., Wang, J., Tang, S., Wang,
Y., & Deng, X. (2018). Magnolol restores the activity of meropenem
against NDM-1-producing Escherichia coli by inhibiting the activity of
metallo-beta-lactamase. Cell Death Discovery,4(28). https://doi.org/10.
1038/s41420-018-0029-6
Mendonc¸a, N., Leit~
ao, J., Manageiro, V., Ferreira, E., & Canic¸a, M. (2007).
Spread of extended-spectrum b-lactamase CTX-M-producing
Escherichia coli clinical isolates in community and nosocomial envi-
ronments in Portugal. Antimicrobial Agents and Chemotherapy,51(6),
1946–1955. https://doi.org/10.1128/AAC.01412-06
Mengs, U., Torsten Pohl, R., & Mitchell, T. (2012). LegalonV
RSIL: The anti-
dote of choice in patients with acute hepatotoxicity from amatoxin
poisoning. Current Pharmaceutical Biotechnology,13(10), 1964–1970.
https://doi.org/10.2174/138920112802273353
Mogana, R., Adhikari, A., Tzar, M. N., Ramliza, R., & Wiart, C. (2020).
Antibacterial activities of the extracts, fractions and isolated com-
pounds from Canarium patentinervium Miq. against bacterial clinical
isolates. BMC Complementary and Alternative Medicine,20(55), 1–11.
Mohanraj, K., Karthikeyan, B. S., Vivek-Ananth, R., P., Chand, R. P. B.,
Aparna, S. R., Mangalapandi, P., & Samal, A. (2018). IMPPAT: A curated
database of Indian Medicinal Plants. Phytochemistry and Therapeutics.
Scientific Reports,8(1), 1–17. https://doi.org/10.1038/s41598-018-
22631-z
Morris, G. M., Ruth, H., Lindstrom, W., Sanner, M. F., Belew, R. K.,
Goodsell, D. S., & Olson, A. J. (2009). Software news and updates
AutoDock4 and AutoDockTools4: Automated docking with selective
receptor flexibility. Journal of Computational Chemistry,30(16), 2785–
2791. https://doi.org/10.1002/jcc.21256
Nadimpalli, M. L., Chan, C. W., & Doron, S. (2021). Antibiotic resistance: A
call to action to prevent the next epidemic of inequality. Nature
Medicine,27(2), 186–187. https://doi.org/10.1038/s41591-021-01229-5
Nichols, D. A., Jaishankar, P., Larson, W., Smith, E., Liu, G., Beyrouthy, R.,
Bonnet, R., Renslo, A. R., & Chen, Y. (2012). Structure-based design of
potent and ligand-efficient inhibitors of CTX-M class A b-lactamase.
Journal of Medicinal Chemistry,55(5), 2163–2172. https://doi.org/10.
1021/jm2014138
Ohene-Agyei, T., Mowla, R., Rahman, T., & Venter, H. (2014).
Phytochemicals increase the antibacterial activity of antibiotics by act-
ing on a drug efflux pump. MicrobiologyOpen,3(6), 885–896. https://
doi.org/10.1002/mbo3.212
Pandey, N., & Cascella, M. (2022). Beta lactam antibiotics. In StatPearls.
StatPearls Publishing. PMID 31424895
Peirano, G., & Pitout, J. D. D. (2019). Extended-spectrum b-lactamase-pro-
ducing enterobacteriaceae: Update on molecular epidemiology and
treatment options. Drugs,79(14), 1529–1541. https://doi.org/10.1007/
s40265-019-01180-3
Pires, D. E. V., Blundell, T. L., & Ascher, D. B. (2015). pkCSM: Predicting
small-molecule pharmacokinetic and toxicity properties using graph-
based signatures. Journal of Medicinal Chemistry,58(9), 4066–4072.
https://doi.org/10.1021/acs.jmedchem.5b00104
Pronk, S., P
all, S., Schulz, R., Larsson, P., Bjelkmar, P., Apostolov, R., Shirts,
M. R., Smith, J. C., Kasson, P. M., Van Der Spoel, D., Hess, B., & Lindahl,
E. (2013). GROMACS 4.5: A high-throughput and highly parallel open
source molecular simulation toolkit. Bioinformatics,29(7), 845–854.
https://doi.org/10.1093/bioinformatics/btt055
Rahman, M., & Khan, K. A. (2019). In silico based unraveling of New Delhi
metallo- b- lactamase (NDM-1) inhibitors from natural compounds: A
molecular docking and molecular dynamics simulation study. Journal
of Biomolecular Structure and Dynamics,38(7), 1–11. https://doi.org/10.
1080/07391102.2019.1627248
Ramdani-Bouguessa, N., Mendonc¸a, N., Leit~
ao, J., Ferreira, E., Tazir, M., &
Canic¸a, M. (2006). CTX-M-3 and CTX-M-15 extended-spectrum b-lacta-
mases in isolates of Escherichia coli from a hospital in Algiers. Journal
of Clinical Microbiology,44(12), 4584–4586. https://doi.org/10.1128/
JCM.01445-06
Remmert, M., Biegert, A., Hauser, A., & S€
oding, J. (2012). HHblits:
Lightning-fast iterative protein sequence searching by HMM-HMM
alignment. Nature Methods,9(2), 173–175. https://doi.org/10.1038/
nmeth.1818
Rhodes, G. (2006). Crystallography made crystal clear (3rd ed.) In
Academic Press. https://doi.org/10.1016/b978-012587073-3/50011-8
Rossolini, G. M., D’Andrea, M. M., & Mugnaioli, C. (2008). The spread of
CTX-M-type extended-spectrum b-lactamases. Clinical Microbiology
and Infection,14(SUPPL. 1), 33–41. https://doi.org/10.1111/j.1469-0691.
2007.01867.x
Song, J.-H., Kim, K.-J., & Lee, B.-Y. (2016). Silibinin suppresses mediators
of inflammation through the inhibition of TLR4-TAK1 pathway in LPS-
induced. Journal of Food and Nutrition Research,4(8), 515–521.
https://doi.org/10.12691/jfnr-4-8-5
Toudji, G. A., Thiombiano, E., Karou, S. D., Anani, K., Adjrah, Y., Gbekley,
H. E., Kiendrebeogo, M., Ameyapoh, Y., & Simpore, J. (2017).
Antibacterial and anti-inflammatory activities of crude extracts of
three Togolese medicinal plants against ESBL Klebsiella pneumoniae
strains. African Journal of Traditional, Complementary and Alternative
Medicines,15(1), 42–58. https://doi.org/10.21010/ajtcam.vi15.1.5
Trott, O., & Olson, A. J. (2009). AutoDock Vina: Improving the speed and
accuracy of docking with a new scoring function, efficient optimiza-
tion, and multithreading. Journal of Computational Chemistry,31(2).
https://doi.org/10.1002/jcc.21334
University of Oxford. (2022). An estimated 1.2 million people died in 2019
from antibiotic-resistant bacterial infections. University of Oxford, pp.
20–22.
16 B. KAR ET AL.

Ur Rahman, S., Ali, T., Ali, I., Khan, N. A., Han, B., & Gao, J. (2018). The
growing genetic and functional diversity of extended spectrum beta-
lactamases. BioMed Research International,2018, 9519718. https://doi.
org/10.1155/2018/9519718
US Department of Health and Human Services, & CDC. (2019). Antibiotic
resistance threats in the United States. Centers for Disease Control and
Prevention, pp. 1–113. https://www.cdc.gov/drugresistance/biggest_
threats.html
Wang, E., Sun, H., Wang, J., Wang, Z., Liu, H., Zhang, J. Z. H., & Hou, T.
(2019). End-point binding free energy calculation with MM/PBSA and
MM/GBSA: Strategies and applications in drug design. Chemical
Reviews,119(16), 9478–9508. https://doi.org/10.1021/acs.chemrev.
9b00055
Wang, X., Yang, Y., Gao, Y., & Niu, X. (2020). Discovery of the novel
inhibitor against New Delhi Metallo- b-lactamase based on virtual
screening and molecular modelling. International Journal of Molecular
Sciences,21(10), 3567.
WHO. (2019). WHO Global report on traditional and complementary medi-
cine 2019. World Health Organization. https://apps.who.int/iris/bit-
stream/handle/10665/312342/9789241515436-eng.pdf?ua=1
Zeynudin, A., Pritsch, M., Schubert, S., Messerer, M., Liegl, G., Hoelscher,
M., Belachew, T., & Wieser, A. (2018). Prevalence and antibiotic sus-
ceptibility pattern of CTX-M type extended-spectrum b-lactamases
among clinical isolates of gram-negative bacilli in Jimma, Ethiopia.
BMC Infectious Diseases,18(1), 1–10. https://doi.org/10.1186/s12879-
018-3436-7
JOURNAL OF BIOMOLECULAR STRUCTURE AND DYNAMICS 17