Content uploaded by Ravi Kant Pathak
Author content
All content in this area was uploaded by Ravi Kant Pathak on Apr 07, 2022
Content may be subject to copyright.
ISSN 0973-2063 (online) 0973-8894 (print)
©Biomedical Informatics (2022)
Bioinformation 18(3): 170-179 (2022)
170
www.bioinformation.net
Research Article
Volume 18(3)
Received February 24, 2022; Revised March 24, 2022; Accepted March 31, 2022, Published March 31, 2022 DOI: 10.6026/97320630018170
Declaration on Publication Ethics:
The author’s state that they adhere with COPE guidelines on publishing ethics as described elsewhere at https://publicationethics.org/.
The authors also undertake that they are not associated with any other third party (governmental or non-governmental agencies) linking
with any form of unethical issues connecting to this publication. The authors also declare that they are not withholding any information
that is misleading to the publisher in regard to this article.
Declaration on official E-mail:
The corresponding author declares that lifetime official e-mail from their institution is not available for all authors
License statement:
This is an Open Access article which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is
properly credited. This is distributed under the terms of the Creative Commons Attribution License
Comments from readers:
Articles published in BIOINFORMATION are open for relevant post publication comments and criticisms, which will be published
immediately linking to the original article without open access charges. Comments should be concise, coherent and critical in less than 1000
words.
Edited by P Kangueane
Citation: Ahmad et al. Bioinformation 18(3): 170-179 (2022)
Molecular dynamics simulation and docking analysis
of NF-κB protein binding with sulindac acid
Shaban Ahmad1, 2,#, Piyush Bhanu3,#, Jitendra Kumar4,#, Ravi Kant Pathak5, Dharmendra Mallick6,
Akshay Uttarkar7, Vidya Niranjan8 & Vachaspati Mishra9,*
1International Centre for Genetic Engineering and Biotechnology, Aruna Asaf Ali Marg, New Delhi 110067, India; 2Department of
Computer Science, Jamia Milia Islamia, New Delhi 110025, India; 3Xome Life Sciences, Bangalore Bioinnovation Centre, Helix Biotech Park,
Bengaluru 560100, Karnataka, India; 4Bangalore Bioinnovation Centre (BBC), Helix Biotech Park, Electronics City Phase 1, Bengaluru
560100, Karnataka, India; 5School of Bioengineering and Biosciences, Lovely Professional University, Jalandhar—Delhi Grand Trunk Rd,
Phagwara 144001, Punjab, India; 6Department of Botany, Deshbandhu College, University of Delhi, Delhi 110019, India; 7Department of
Biotechnology, RV College of Engineering, RV Vidyanikethan Post, Mysuru Road, Bengaluru 560059, India; 8Department of
Biotechnology, RV College of Engineering, RV Vidyanikethan Post, Mysuru Road, Bengaluru 560059, India; 9Department of Botany,
Hindu College, University of Delhi, Delhi 110007, India;*Correspondence, #equal contribution.
Author contacts:
Shaban Ahmad - E-mail: shaban.12082000@gmail.com
Piyush Bhanu - E-mail: pb.inbusiness@gmail.com
Jitendra Kumar - E-mail: director@bioinnovationcentre.com
ISSN 0973-2063 (online) 0973-8894 (print)
©Biomedical Informatics (2022)
Bioinformation 18(3): 170-179 (2022)
171
Ravi Kant Pathak - E-mail: ravibt36@gmail.com
AkshayUttarkar - E-mail: akshayuc@rvce.edu.in
Vidya Niranjan - E-mail: vidya.n@rvce.edu.in
Vachaspati Mishra - E-mail: mishravachaspati31@gmail.com
Abstract:
It is of interest to document the Molecular Dynamics Simulation and docking analysis of NF-κB target with sulindac sodium in combating
COVID-19 for further consideration. Sulindac is a nonsteroidal anti-inflammatory drug (NSAID) of the arylalkanoic acid class that is
marketed by Merck under the brand name Clinoril. We show the binding features of sulindac sodium with NF-κB that can be useful in
drug repurposing in COVID-19 therapy.
Keywords: NF-κB; molecular dynamics simulation; COVID-19; SARS-CoV2; sulindac sodium
Background:
Sulindac is a non-steroidal anti-inflammatory agent (NSAIA) that
can block nuclear factor-κB (NF-κB), a transcription factor (TF)
located primarily in the cytoplasm in a complex while in an inactive
state. Activation of this TF makes it transition from a latent
cytoplasmic form to a nuclear DNA binding state [1] and has been
demonstrated to have a role in the prevention of colon cancer [2] by
its inhibition [1, 3].This TF is known to selectively inhibit
interferon-gamma-induced expression of the chemokine CXCL9
gene in mouse macrophages. [4] The growth inhibitory and anti-
inflammatory properties of sulindac are possibly due to its ability
to reduce prostaglandin synthesis by cyclooxygenase inhibition [2].
Furthermore, Yamamoto et al. [2] have demonstrated that aspirin
and sodium salicylate inhibited the activity of an IκB kinase β
(IKKβ) that is required to activate the NF-κB pathway, which leads
to nuclear translocation of NF-κB, where it binds to its cognate
DNA and activates transcription of a wide variety of genes
involved in host immunity, inflammation, cell proliferation and
apoptosis [2, 3]. Various NF-κB inducers are known to be highly
variable and include aggregates of bacterial lipopolysaccharides,
ionizing radiation, reactive oxygen species (ROS), cytokines, such
as tumour necrosis factor alpha (TNF-α) and interleukin 1-beta (IL-
1β) and viral DNA and RNA [4]. Upon activation, NF-κB promotes
the gene expression of a broad range of cytokines (e.g., IL-1, IL-2,
IL-6, IL-12, TNF-α, LT-α, LT-β, and GM-CSF), chemokines (e.g., IL-
8, MIP-1, MCP1, RANTES, and eotaxin), adhesion molecules (e.g.,
ICAM, VCAM, and E-selectin), acute phase proteins (e.g., serum
amyloid A: SAA), and inducible effector enzymes (e.g., inducible
nitric oxide synthase: iNOS and cyclooxygenase-2: COX-2). The TF
NF-κB can be considered as a “quick action” primary transcription
factor that is able to regulate myriad of cellular responses,
including the host’s early innate immune response to infection, and
is also associated with chronic inflammatory states, viral infections,
septic shock syndrome and multiorgan failure [5, 6] viz-a-viz
ascribed as an anti-apoptotic TF [7].The constitutive activation of
NF-κB pathways has been responsible for stimulation of
inflammatory diseases, such as multiple sclerosis and rheumatoid
arthritis [8, 9]. Furthermore, p38 MAPK-based activation of NF-kB
has also been reported [9]. Drug induced suppression of NF-kB has
been found to have a strong potential in proapoptotic signal
modulation therapy [10].
Recently, a study utilizing molecular dynamics simulations (MDS) -
based analysis targeting human angiotensin-converting enzyme 2
(hACE2) as a receptor showed interesting results on development
of a lead candidate to be used later as a therapeutic drug against
COVID-19 [11] Though in the current scenario, the spread of
COVID 19-based pandemic has slowed down to a great extent, but
its resurgence cannot be ruled out. Due to the lack of adequate
specific treatments for COVID-19, there is an urgency to develop or
repurpose drugs to help end the epidemic completely. The drug
discovery process has received a great impetus with the emergence
of computer-aided drug discovery (CADD) and has been
instrumental over the last decade in exploring protein inhibitors in
protein-drug interactions and protein–protein interactions [12].
Since the process involved in the development of a candidate drug
into an approved drug is lengthy and expensive, a combination of
computational methodologies such as virtual screening, docking,
molecular dynamics simulation, and binding free energy evaluation
contributes to identifying potential drug candidates from
compound libraries [13].To date, MDS and structure-based virtual
screening studies have been carried out to understand the SARS-
CoV-2 spike protein functions [14-16], role of repurposed protease
inhibitors in blocking SARS-CoV-2 [17], effects of some antiviral
drugs on SARS-CoV-2 [18], effect of temperature on the structure of
the spike protein [19], SARS-CoV-2-targetted potent drugs that can
be effective in controlling COVID-19 [20, 21] as well as many other
aspects of biology qualifying to control SARS-CoV-2 progression.
However, the rate of occurrence of different pathological human
phenotypes and their heterogeneous lethality rates arising from
constant mutations of SARS-CoV-2 [22, 23] indicate serious
problems that need to be appropriately dealt with. A paradigm
shift in the strategy on targeting COVID-19 borne biomolecules-for
therapeutic purposes to human-based molecular biomarkers is
much necessitated in the current scenario, since the current
approaches based on former fail to yield appreciable outcomes. A
recent study revealed the importance of the NF-κB pathway in
developing therapy regimens for critical COVID-19 patients [24].
Other reports have also substantiated the above fact by arguing that
therapeutic benefits of NF-κB inhibitors, including dexamethasone,
a synthetic form of glucocorticoid, have increasingly been realized
now [17]. Abnormal activation of NF-κB resulting from SARS-CoV-
2 infection might be associated with the pathogenic profile of
immune cells, cytokine storms and multiorgan defects [23].
Therefore, we currently explored the structural details of NF-κB via
ISSN 0973-2063 (online) 0973-8894 (print)
©Biomedical Informatics (2022)
Bioinformation 18(3): 170-179 (2022)
172
an in silico approach wherein we pinpointed the protein pockets for
the binding of SARS-CoV-2 spike proteins. A hypothetical
explanation currently generated is that upon entering the human
body, the COVID-19 viral protein might be triggering the action of
the NF-κB directly or through activation of IKK, the kinase
phosphorylating IκBs that otherwise remain bound with NF-κB to
keep it in an inactive state in the cytoplasm. This argument is made
in the light of the fact that another complex member, IκBα is known
to be phosphorylated by specific kinases at two sites near the N-
terminus (Ser-32 and Ser-36), while the phosphorylated protein is
then ubiquitinated at Lys-21 and Lys-22, leading to proteosome-
mediated degradation [24] and release of NF-κB from the complex.
The removal of IκB unmasks the nuclear localization sequence
(NLS) of NF-κB and allows its translocation to the nucleus [24].
Thus, IκBα is a multifunctional inhibitor of NF-κB that blocks
nuclear translocation, DNA binding, and phosphorylation by
protein kinase A (PKA). Inhibition of NF-kB blocks the NF-κB-
mediated deactivation of immune and inflammatory responses to
stimuli, such as cytokines or bacterial/viral infection products [24].
Along with NF-κB, sulindac has been of interest because of its
many roles that includes also as a chemo preventive agent for
adenomatous colorectal polyps and colon cancer [25] We show here
that sulindac is able to bind specifically to a pocket on NF-κB, a
concept that becomes theoretically important for exploring its
therapeutic potential [26- 29].
Figure 1: Root mean square deviation (RMSD) of the protein and
ligand after the initial RMSD values were stabilized. This plot
shows RMSD values for the protein on the left Y-axis, whereas for
the ligand, these values are indicated on the right Y-axis. The
RMSD graph for c-αis shown in blue colour, and for ligand fit on
the protein, it is in red colour.
Materials and Methods:
Protein and Ligand Preparation:
Coordinates of the X-ray diffraction-based crystal structure of the I-
κ-B-α/NF-κB complex with solvents (PDB ID: 1NFI) at a resolution
of 2.70 Å were downloaded from the Protein Data Bank (PDB).
Solvent molecules and chains A, B, and F were removed during
protein preparation using the Protein Preparation wizard from
Schrodinger Maestro (Schrodinger, LLC, and New York, NY, USA).
The remaining structures were processed using the Protein
Preparation wizard using appropriate methodologies. The SARS-
CoV-2 structure possessing a loop (residues 376–381) was missing
in the PDB structure and hence was modelled using Schrödinger’s
Prime module. Hydrogen atoms were incorporated, and a standard
protonation state at pH 7 was used. Bond orders were assigned
using the chemical components dictionary (CCD) database. The
heterostate was generated using Epik with a pH of 7 (±2.0), and
Prime was used to fill the missing chains and loops. While refining
the structure, PROPKA with pH 7.0 was used along with the
sample water orientation. While minimizing the structure, a root
mean square deviation (RMSD) of 0.30 Å with the OPLS_2005 force
field was used. The ligand was selected based on published
information regarding its use as a therapeutic molecule for various
human diseases. The ligand was prepared using the Ligprep
wizard of Schrodinger with the OPLS_2005 force field; the Epik job
was submitted with the metal-binding site and included the
original state. Thirty-two stereoisomers were assigned per ligand
with specified chirality’s. Glide was used to filter the search to
locate the ligand in the active-site region of the receptor. The shape
and properties of the receptor were represented on a grid to
provide a more accurate scoring of the ligand poses. The docked
complexes were superimposed on the original crystal structure to
calculate the RMSD (Figure 1).
Virtual Screening and Molecular Docking:
The Glide grid generator was used for generating the grid for blind
docking. A Schrodinger virtual screening workflow (VSW) was
used to score the virtual screening with default parameters
employing the Glide program of Schrödinger. Use of HTVS mode
allowed the elimination of most of the stereoisomers, and only 10%
of the total ligands could be retained that passed the screening.
Ligands following QikProp and Lipinski’s rule were filtered for
docking. Docking in the VSW was performed with Epik state
penalties and further passed through HTVS, SP and XP docking
modes. The interactions of the selected ligand and protein docked
complexes were analyzed by a pose viewer.
Molecular Dynamics Simulation:
To study the dynamic behavior of the protein complex under
simulated physiological conditions, MDSs of the protein-ligand
(PL) complex were performed using Desmond, which is available
with Schrodinger Maestro (v12.5). The PL complex (9873 atoms)
was solvated in a 10 × 10 × 10 Å orthorhombic periodic box built
with TIP3P water molecules [30]. The whole system was
neutralized by adding an appropriate number of 6 Na+ counterions.
This solvated system was energy minimized and position
restrained with OPLS_2005 as the force field. Furthermore, 100 ns of
MDS was carried out at 1 atm pressure and 300 K temperature,
implementing the NPT ensemble with a recording interval of 100
ps, resulting in a total of 1000 read frames. Finally, various
parameters of the MDS, such as ligand-binding site analysis,
RMSD, root mean square fluctuation (RMSF), PL contacts,
secondary structure element (SSE) analysis, etc., were also analyzed
to check the stability, compactness, structural fluctuations and PL
interactions in the solvated system (Figure 2).
ISSN 0973-2063 (online) 0973-8894 (print)
©Biomedical Informatics (2022)
Bioinformation 18(3): 170-179 (2022)
173
Interaction Analyses
Structural Data:
The PDB was used to extract all 3-D structural information. The X-
ray diffraction-based crystal structure of the I-κ-B-α/NF-κB
complex with solvents (PDB ID: 1NFI) at a resolution of 2.70 Å was
used for this study. Solvents and chains A, B, and F were removed
during protein preparation. The structure of the SARS-CoV-2 spike
glycoprotein (closed state) solved using electron microscopy (PDB
ID: 6VXX) with a resolution of 2.80 Å was obtained. Furthermore,
all water molecules were removed from both structural data sets.
The educational version of PyMOL (The PyMOL Molecular
Graphics System, v1.2r3pre, Schrödinger, LLC) was used to
generate the images derived to understand and analyse the
structure and inter chain interaction information.
Molecular Docking:
The ligand sulindac chosen in the current study for the mentioned
reasons has its docking score and binding affinity shown in Figure
3. The complex of the drug molecule docked in the pocket of NF-κB
was designated S0 (Figure 4A) and was taken as the reference for
further interaction analysis.
Pocket Analysis:
The molecule NF-κB was found to having multiple pockets, as
calculated using the CASTp 3.0 server [30, 31] Three largest pockets
(based on the accessible area) of these were considered for analysis.
The residues present in these pockets interacted with the spike
protein on exposure changing the area of the pockets significantly
due to this interaction and also upon the drug interaction or the
combination of the spike protein and the drug together.
Protein-Protein Docking:
Protein-protein docking studies were carried out using ClusPro 2.0
[32] with minor changes following the protocols used in Jorgensen
et al. [33]. This job was carried out to understand the interaction of
the SARS-CoV-2 spike glycoprotein with NF-κB in the presence and
absence of the drug molecule. The best pose of the spike protein
binding with the NF-κB-IkB complex was selected based on the
cluster size, which is how the models are ranked in ClusPro. S1
(Figure 4B) represents the complex of NF-κB with the SARS-CoV-2
spike glycoprotein, and S2 (Figure 4C) represents the complex of S0
with the SARS-CoV-2 spike glycoprotein.
LigPlot+ v.2.2 [34] was used to find the ligand-protein interaction of
complex S0 (Figure 2). PDBsum was used to calculate the
interaction between NF-κB and IkB (Figure 5A, 5B), the interaction
between the spike protein and NF-κB (Figure 5B) and the
interaction between the spike protein and NF-κB docked with drug
molecules. PyMOL was used to analyse the structural differences
imposed on NF-κB when it interacted with the drug molecule and
the spike protein in the presence of the drug molecule.
Figure 2: RMSF of the protein backbone and ligand complex. Red
colour shows the B factor means of the PDB structure, and green
colour indicates the interaction of the ligand with the protein.
MM-GBSA Studies:
Prime MM-GBSA (molecular mechanics, the generalized born
model and solvent accessibility method) calculations were
performed with the docked complex of receptor and sulindac. The
implicit solvent model used is VSGB [35] which has an efficient
approximation of the solvation free energy with Surface
Generalized Born (SGB) over Vacuum and chloroform solvation
models. The following calculations were made
1 Rec Strain = Receptor (from optimized complex) − Receptor
2 Lig Strain = Ligand (from optimized complex) − Ligand
3 MMGBSA dG Bind = Complex − Receptor − Ligand
4 MMGBSA dG Bind (NS) = Complex – Receptor (from optimized complex) –
Ligand (from optimized complex) = MMGBSA dG Bind
Rec Strain - Lig Strain:
NS here refers to binding or interaction energy without any
involvement of conformational changes occurring in the receptor or
ligand. The potential energy of the complex will be reported in
kcal/mol.
ISSN 0973-2063 (online) 0973-8894 (print)
©Biomedical Informatics (2022)
Bioinformation 18(3): 170-179 (2022)
174
Figure 3: (A) Diagram showing detailed ligand atom interactions
with the protein residues. Interactions that occur more than 7.0% of
the simulation time in the selected trajectory from 0.00 through
100.00 ns. The highlighted structure represents the ligand (drug)
molecule. The interacting residues are shown in circular shapes
with the representation of chains as C (representing residues of the
NFkB molecule) and E (representing residues of IkBα). Circles with
H2O represent the water-mediated H-bond interactions. (B) A
timeline representation of the interactions and contacts (H-bond,
hydrophobic, ionic, and Water Bridge) summarized in the above
figure. The top panel shows the total number of specific contacts
that the protein makes with the ligand over the course of the
trajectory. The bottom panel shows specific residues interacting
with the ligand in each trajectory frame. Some residues make more
than one specific contact with the ligand, which is represented by a
darker shade of orange according to the scale shown to the right of
the plot.
Results and Discussion:
MD Simulation Studies:
The complex RMSD in the initial phase at 1.63 Å went to 4.11 Å
while stabilizing the structure for 100 ns of simulation. Initially, at
up to 50 ns, the RMSD did not fluctuate much, but after that, the
stability fluctuated for 10 ns and stabilized subsequently (Figure 1).
The RMSF plot analysis displayed minimal fluctuations in the
protein structures (Figure 2). Furthermore, minimal fluctuations
were also observed in the PL complex, and the RMSF plot showed
fluctuations in some regions of the protein. The results of the
residue interaction analysis after docking are highlighted with
specific colours based on the self-explanatory features (Figure3).
The docking interactions of sulindac (Figure 4A, 4B) showed that it
docked in the largest pocket of the NF-κB-IkB complex that
contains most of the residues interacting with the spike protein. It is
possible that drug binding to the complex interferes with the above
interaction of the spike protein, however, needs to be ascertained
with more robust wet laboratory-based experimental designs. The
sulindac was found to bind to other pocket than the one occupied
by NF-κB-IkB complex (Figure 4C). Moreover, data shows the
importance of NF-κB as a suitable drug receptor for COVID-19 for
its proven roles as molecular targets in drug discovery-based in
silico studies (Table 1).
The interaction made by the drug molecule with the residues of the
active site cavity of NF-kB, IkB and P50 peptide chains is dynamic
in nature and formed, broke and reformed during the simulation
duration (Figure 3). A brief duration of 20 ns was observed when
no interaction occurred between the drug molecule and the NF-kB
active site residues. Although there was an interaction between the
ligand (drug) molecule and residues of NF-kB (chain C) and IkB
(chain E) in the first 50 ns, the interaction between the drug and IkB
(chain E) disappeared after 50 ns. This loss of interaction
strengthened the ligand-NFkB interaction.
Table 1: The binding affinities of various drugs with NF-κB are shown. This shows that NF-κB is a candidate receptor in various in silico drug-based interventions that may have
potential for being a therapeutic target.
Sl.
No.
Bioactive compound/ligand
Binding pocket
Outcomes achieved/remarks
References
1
Piperine and piperlongumine
DNA binding sites
Docking results of the analogues had higher affinity towards NF-
κB and IL-1β than natural molecules
[36]
2
71 anti-cancer compounds along with
dipeptidyl boronic acid, betulinic acid, and
glycyrrhetinic acid inhibitors
Active sites: Gly47, Ala49, Thr21, and Thr1 of the receptor
Docking was at par with the drug, Bortezomib against
proteasome target of NF-κB pathway
[37]
3
Compound series designed based on
quinazoline scaffold pharmacophore model
p50 subunit of NF-κB
Compounds can inhit NF-κB function and reduce the
proliferation of numerous tumor cell lines especially compound
2a
[38]
ISSN 0973-2063 (online) 0973-8894 (print)
©Biomedical Informatics (2022)
Bioinformation 18(3): 170-179 (2022)
175
4
P100 peptide
Binding sites of IκB Kinase α(IKK α)
Highlightes the different roles of NIK and IKK in regulating p100
processing and understands the mechanisms that mediate
control of p100 processing
[39]
5
Four series of 24 thienopyrimidine/N-
methylpicolinamide derivatives substituted
with pyrimidine
Active sites of TAK1 kinases
Compound 38 inhibited TAK1 phosphorylation, I κBα
degradation, TNF-α-induced IκBα phosphorylation, , reduce the
expression of p65 and p65 phosphorylation
[40]
6
Growth arrest specific facgtor 6 (Gas6)
Autophosphorylation docking site- Tyr867
Mertk receptor showed disti nct effects for phagocytosis
[41]
7
1, 8-dihydroxy-4-methylanthracene-9, 10-
dione (DHMA) from Luffa acutangal a
DNA-binding region (DBR) of NF-κB
DHMA plays a role in cancer prevention and treatment by
having an altered binding affinities against NF-κB and DNA
[42]
8
Natural compounds found in Cupressus
pyramidalis and Aegle marmelos extracts
DNA binding sites of proteins
These lead molecules down-regulated the expression of IL8 gene
and inhibited the activity of NF-kappaB-50
[43]
9
Phenolic fraction of Citru s maxima
(hesperidin, naringenine, naringin and
dexamethasone)
DNA binding region of NF-κB
Docking studies of the phenolic fraction compounds had an
inhibitory effect against NF-κB
[44]
10
Diphosphorylated 21-mer p100 peptide
model
Four argenine residues Arg285, Arg410, Arg431, Arg521 and
Arg474 in hydrophobe B-TrCP
Peptide p100 disrupts the IκB-α/NF-κB signalling pathway
[45]
11
Small molecule, NSC-127102
Phosphorylation of p65 at serine 276
NSC-127102 hinders IL-8, VCAM-1 expression and
phosphorylation at serine 276
[46]
12
Scopoletin (Pharmacologically active
coumarin)
Keap1, NF-kB, HO-1 and p38MAPk are the targets against
vancomycin nephrotoxicity
Scopoletin helps in protecting renal inury by interferring with
Nrf2/HO-1 and IkBa/p65 NF-kB signalling pathway as well as
resorting the antioxidant defense system
[47]
13
Ergosta-7, 9 (11), 22-trien-3β-ol (EK100)
Interferes with Lipopolysaccharide (LPS) docking to the LPS-
binding protein (LBP), transferred to the cluster of differenti ation 14
(CD14), and bonded to TLR4/myeloid differentiation-2 (MD-2) co-
receptors
EK100 equally attenuated NF-κB gene expression and TLR4/NF-
κB inflammatory pathway in Drosophila
[48]
Figure 4: X-ray diffraction-based crystal structure of the IκB-α/NF-
κB complex with solvents (PDB ID: 1NFI) at resolution of 2.70 Å,
showing various chains in colour formats representing different
molecules employed in the current study. For example, chain C
(shown as a green cartoon model) represents NF-κB-P65, chain D
(shown in a cyan cartoon model) represents NF-κB-P50, and chain
E (shown in a magenta cartoon model) represents IkBα. (A)
Structure of the complex formed after docking of the drug molecule
with NF-κB. The highlighted region represents the binding pocket
of the drug molecule in the NF-κB molecule. The binding pocket is
represented by the labelled residues (G209, D210, E211, D257, R253,
R259, S269, E285, etc., of the NFkB molecule and I102, N182, Q183,
H184, etc., of the IkBα molecule). (B) Structure of the complex
formed after protein-protein docking of the SARS-CoV-2 spike
glycoprotein (closed state) (PDB ID: 6VXX, shown in orange colour
at a resolution of 2.80 Å) with the NF-κB molecule. The highlighted
region represents the binding pocket of the spike protein in the NF-
κB molecule. The binding pocket is represented by the labelled
residues (G209, D210, E211, D257, R253, R259, S269, E285, etc., of
the NFkB molecule and N182, Q183, H184, etc., of the IkBα
molecule). This information indicates that the drug molecule binds
in the same pocket that the spike protein used to interact with NF-
κB, possibly to activate its translocation to the nucleus. (C)
Structure of the complex formed after protein-protein docking of
the SARS-CoV-2 spike glycoprotein (shown in orange colour) with
the NF-κB molecule (already docked with the drug molecule). The
highlighted region represents the binding pocket of the spike
protein in the NF-κB molecule in the presence of the drug molecule.
The drug molecule occupies the same binding pocket on NF-κB
that is also used by the spike protein during docking interaction,
thus inhibiting the spike protein from interacting with NF-κB at its
designated site. In the interim, the spike protein interacts with NF-
κB at a different site and hence may not be able to activate it.
Currently, these assumptions need validating experimental proof,
which is lacking because of the strict restraint on experimentation
with live samples from active human COVID-19 cases.
ISSN 0973-2063 (online) 0973-8894 (print)
©Biomedical Informatics (2022)
Bioinformation 18(3): 170-179 (2022)
176
Figure 5: A 3-D structural alignment of the NF-κB-drug complexes
(represented in a red cartoon model) and spike-protein-NF-κB-drug
complexes (represented in a green cartoon model), with the NF-κB
molecule represented in an orange cartoon model. (A) These
alignments show the conformational changes imposed in the
structure of the NF-κB molecule upon binding of the drug and
spike protein. (B) Sequence-based comparison of the interacting
residues of the NF-κB molecule with the drug sulindac sodium
(yellow coloured bar), IkB molecule (green coloured bar) and spike
protein (red coloured bar).
The changes in the 3-D structure of the NF-κB-IkB complex due to
the binding of the drug molecule and spike protein in the presence
of the drug molecule were compared to that of the unbound
complex (Figure 5A). The binding of the drug molecule to the NF-
κB-IkB complex led to an increase in the compactness of the
complex. The area and volume data of the pockets presented in
Figure 5A corroborated this observation. The binding of the spike
protein in the presence of the drug molecule further increased the
compactness of this complex, possibly due to increased interaction
between the chains of the NF-κB-IkB complex. Additionally, we
made a comparison of the amino acid (aa) residues of the spike
protein, MAP kinase and pLC gamma during their interactions
with the residues of the NF-κB molecule to identify if any common
pattern existing in aa selection by sulindac during these
interactions. However, all three sets of interactions, such as NF-κB-
sulindac docked with spike protein, NF-κB sulindac docked with
MAPK and NF-κB sulindac docked with pLC gamma had different
aa residues interacting with the chosen ligand (sulindac)(Figure 6),
revealing all these as independent events. In our studies, MM-
GBSA studies on docked complexes revealed dG bind values that
were reasonably in agreement with ranking based experimental
scores, i.e., binding values We first calculated the MMGBSA dG
Bind = Prime Energy (Optimized Complex) − Prime Energy
(Optimized Free Ligand) − Prime Energy (Optimized Free
Receptor) and second, MMGBSA dG Bind(NS) = Prime Energy
(Optimized Complex) − Prime Energy (Ligand Geometry From
Optimized Complex) − Prime Energy (Receptor Geometry From
Optimized Complex). The values were found to be −2.15 kcal/mol
for MMGBSA dG bind and −2.71 kcal/mol for MMGBSA dG bind
(NS). The atoms contributing to the binding affinities are shown in
Figure 7.
Figure 6: Comparison of the residues of the spike protein, MAP
kinase and pLC gamma interacting with the residues of the NF-κB
molecule. The underlined residues are in common (black: common
in the s protein and pLC gamma, red: common in the spike protein
and MAP kinase, and blue: common in all three).
Our understanding on the involvement of NF-κB signalling
pathway in COVID-19 is limited, but NF-κB inhibitors have been
increasingly utilized to gain therapeutic benefits in many human
diseases [49, 50]. Abnormal activation of NF-κB in the manifestation
of SARS-CoV-2 infection is reported for the pathogenic profile of
immune cells, cytokine storms and multi organ defects and
cytokine storm alone is capable of triggering excessive
inflammatory response to SARS-CoV-2 and is extremely
responsible for disease severity [51]. Thus, pharmacological
inactivation of the NF-κB, such as proposed currently in the use of
sulindac, NF-κB-kinase interaction either directly or by inhibiting
the dephosphorylation of its inhibitors, such as IKKα, IKKβ and
other associated molecules, strongly represents a potential
ISSN 0973-2063 (online) 0973-8894 (print)
©Biomedical Informatics (2022)
Bioinformation 18(3): 170-179 (2022)
177
therapeutic target to treat the symptoms of COVID-19 in clinical
trials. A survey of literature since the emergence of COVID-19
clearly reveals that majority of the drugs formulated to tackle
SARS-CoV-2 thus far, have been structured to target virus [52].
However, given the current situation where the virus mutates so
rapidly, our conceptual paradigm on targeting the key viral
proteins through drug intervention need to get changed towards
the host proteins, such as human proteins that hold key portfolio in
SARS-CoV-2 – mediated infections. This would possibly take care
of the various mutants of SARS-CoV-2 that have been difficult to
control until now. Furthermore, cytokine storm in COVID-19 [53]
and various drugs that were initially used for therapy [54] need to
be revisited in light of modern researches based on drug
repurposing, which are potentially very effective in treatments of
even the anomalies arising post-COVID-19 infection and can thus
counteract the many aberrant responses thereof.
Figure 7: MM-GBSA energy based visualization of docked complex
highlighting the docked complex of receptor and sulindac sodium
are represented with colour highlights on the ligand atoms
showing contributions towards binding affinity. Blue represents the
receptor and adaptive colors on ligand from Red to Green
showcasing higher negative values to least values, respectively. The
yellow dashed line represents the hydrogen bond with the
receptor.
Author contributions:
S.A.: MDS analysis, Original Draft; P.B.: Docking analysis and
Report; J.K.: Docking analysis supervision, Review and Editing;
R.P.: MDS analysis confirmation and Report; D.M.: Review and
Editing; A.U.: MM-GBSA analysis; V.N.: MM-GBSA report; V.M.:
Conceptualization, Revision and Editing, Conclusion and
Discussion, Formatting the MS, Overseeing the project and the MS
Funding:
The funding acquisition was made from the Bangalore Bio
Innovation Centre, Karanataka Innovation and Technology Society,
Department of Electronics, IT, BT and S&T, Government of
Karnataka, India, towards paying the publication cost.
Acknowledgments:
The authors acknowledge the International Centre for Genetic
Engineering and Biotechnology, New Delhi, for providing facilities
for MDS studies and the Bangalore Bio Innovation Centre,
Department of Electronics, IT, BT and S&T, Government of
Karnataka, India, for funding acquisition towards paying the
publication cost.
Conflicts of interest: The authors declare that there is no conflict of
interest.
References:
[1] Housby JN et al. Cytokine. 1999 11:347. [PMID: 10328874]
[2] Yamamoto Y et al. J Biol Chem. 1999 274:27307. [PMID:
10480951]
[3] Sakaeda Y et al. Biochem Biophys Res Commun. 2006
350:339. [PMID: 17010317]
[4] Tegeder I et al. FASEB J. 2001 15:2057. [PMID: 11641233]
[5] Beg AA et al. Genes Dev. 1992 6:1899. [PMID: 1340770]
[6] Oeckinghaus A & Ghosh S Cold Spring Harb Perspect Biol.
2009 1:a000034. [PMID: 20066092]
[7] Sakaeda Y et al. Biochem Biophys Res Commun. 2006
350:339. [PMID: 17010317]
[8] Li Q & Verma IM Nat. Rev. Immunol. 2002 2:725. [PMID:
12360211]
[9] Liu T et al. Signal Transduct Target Ther. 2017 2:17023.
[PMID: 29158945]
[10] Washo-Stultz D et al. Cancer Lett. 2002 177:129. [PMID:
11825660]
[11] Neufeldt CJ et al. Commun Biol. 2022 5:45. [PMID:
35022513]
[12] Bradbury CM et al. Cancer Res. 2001 61:7689. [PMID:
11606413].
[13] Keretsu S et al. Sci. Rep.2020 10:17716. [PMID: 33077821]
[14] Elmezayen AD et al. J Biomol Struct Dyn. 2020 39:2980.
[PMID: 32306862]
[15] McCarty MF et al. Integr Cancer Ther. 2006 5:252. [PMID:
16880431]
[16] Pablo R Arantes et al. ACS Cent. Sci. 2020 6:1654. [PMID:
33140032]
[17] Al-Karmalawy AA et al. Front Chem. 2021 9:661230. [PMID:
34017819]
[18] Selvaraj C et al. J. Biomol. Struct. Dyn. 2020 39:4582. [PMID:
32567979]
[19] Cardoso WB & Mendanha S J. Mol. Struct. 2021 1225:129.
[PMID: 32863430]
[20] Rath SL & Kumar K Front Mol Biosci. 2020 7:583. [PMID:
33195427]
[21] Lokhande KB et al. J Biomol Struct Dyn.2020 1:12. [PMID:
32815481]
[22] Justo A et al. Sci Rep. 2021 11:17755. [PMID: 34493762]
[23] Hariharan A et al. Inflammopharmacology 2021 29:91.
[PMID: 33159646]
[24] Catanzaro M. Signal Transduct Target Ther. 2020 5:84.
[PMID: 32467561]
[25] Wang X et al. Front. Cell Dev. Biol.2020 8: 574706. [PMID:
33224945]
[26] Wang Q et al. Cell 2020 894:904. [PMID: 32275855]
[27] https://go.drugbank.com/drugs/DB00605
[28] Brunell D et al. Drug Metab. Dispos. 2011 1014:1021. [PMID:
21383205]
ISSN 0973-2063 (online) 0973-8894 (print)
©Biomedical Informatics (2022)
Bioinformation 18(3): 170-179 (2022)
178
[29] Marchetti M et al. PLoS ONE 2009 4:e5804. [PMID:
19503837]
[30] Binkowski TA et al. Nucleic Acids Res. 2003 31:3352. [PMID:
12824325]
[31] Tian W et al. Nucleic Acids Res. 2018 46:W363. [PMID:
29860391]
[32] Kozakov D et al. Nat Protoc. 2017 12:255. [PMID: 28079879]
[33] Thompson PA et al. Clin Cancer Res. 2021 27:5660. [PMID:
34112707]
[34] Laskowski RA & Swindells MB J. Chem. Inf. Model.2011
51:2778. [PMID: 21919503]
[35] Li J et al. Proteins. 2011 79:2794. [PMID: 21905107]
[36] Zazeri G et al. Molecules. 2020 25:2841 [PMID: 32575582]
[37] Yadav D et al. J Biomol Struct Dyn. 2020 38:3621. [PMID:
31514715]
[38] Xu L & Russu WA Bioorg Med Chem. 2013 21:540. [PMID:
23219854]
[39] Xiao G et al. J Biol Chem. 2004 279:30099. [PMID: 15140882]
[40] Wang L et al. Eur J Med Chem. 2021 223:113576 [PMID:
34153577]
[41] Tibrewal N et al. J Biol Chem. 2008 283:3618. [PMID:
18039660]
[42] Ramar V & Pappu S Comput Biol Chem. 2016 29:35. [PMID:
27061144]
[43] Piccagli L et al. BMC Struct Biol. 2008 8:38. [PMID:
18768082]
[44] Nandeesh R et al. Biomed Pharmacother. 2018 108:1535.
[PMID: 30372855]
[45] Melikian M et al. J Chem Inf Model. 2017 57:223.
[PMID:28004927]
[46] Law M et al. Cancer Lett. 2010 291:217. [PMID: 19910110].
[47] Khalaf MM et al. Int Immunopharmacol. 2022 102:108382.
[PMID: 34848155]
[48] Hsieh WT et al. Int J Mol Sci. 2021 22:6511. [PMID:
34204506]
[49] Li J et al. Proteins 2011 79:2794. [PMID: 21905107]
[50] Park M H & Hong JT Cells 2016 5:15. [PMID: 27043634]
[51] Yu H et al. Signal Transduct. Target Ther. 2020 5:209. [PMID:
32958760]
[52] Castelli V et al. Frontiers in Immunology 2020 11:2132.
[PMID: 32983172]
[53] McKee DL et al. Pharmacological Research 2020 157:104859
[PMID: 32360480]
ISSN 0973-2063 (online) 0973-8894 (print)
©Biomedical Informatics (2022)
Bioinformation 18(3): 170-179 (2022)
179