Identification of thrombin inhibiting antithrombin-III like protein from Punica granatum using in silico approach and in vitro validation of thrombin inhibition activity in crude protein

Abstract Thrombosis is characterized by the formation of clots in the blood vessels. Antithrombin-III deficiency in the blood causes thrombus formation. Supplementing antithrombin-III may serve as anticoagulant therapy. In the present studies, an antithrombin like Protein from Punica granatum has been identified and characterized using in silico approach. Based on sequence homology, an ALPP was selected depending upon its highest binding affinity of −41.28 kcal/mol with thrombin. Thrombin structure complexed with ALPP was docked with TAME using AutoDock Vina. No binding was observed for TAME at Ser195 of thrombin. MD simulation (50 ns) was performed to evaluate the flexibility and stability of docked complexes. In vitro assays with crude protein showed 78% thrombin inhibition at 5 µg and calculated IC50 value was 0.188 µg. The presence of thrombin inhibitors in crude protein was also confirmed by reverse zymography. Thus, it is very likely that the protein identified from P. granatum may act as thrombin inhibitor. Graphical Abstract


Introduction
Thrombosis/thromboembolic disorder is characterized as the formation of the blood clot inside an injured blood vessel and obstructing the flow of blood through the circulatory system (Sevitt 1970).Excessive thrombin formation results in a deficiency of the natural thrombin inhibitors of the procoagulant system (Pabinger and Schneider 1996).The probability of thrombosis to occur is approximately 70-90% at the age of 50-60 years (Pabinger and Schneider 1996).Based on the interactions with thrombinantithrombin have been categorized as bivalent DTI that binds with both the active sites and exosite 1 of thrombin (e.g.Hirudin, Desirudin, and Lepidurin), univalent DTIs that binds with only active site of thrombin (e.g.Argatroban, Inogatran, Melagatran, and Dabigatran) and allosteric inhibitors that regulates the thrombin by binding to either exosite 1 or 2 and Na þ binding site (Weitz and Bates 2003).Anticoagulants like Warfarin, Vitamin K are generally given orally and heparin is administered via injection (McRae et al. 2021).Both are associated with various side effects and moreover, administration of these drugs require self-monitoring and self-management (Veiraiah and Routledge 2009).
In blood thrombin may act as procoagulant and anticoagulant (Narayanan 1999).As procoagulant thrombin mediates the conversion of fibrinogen to fibrin and activation of clotting factors (V, VIII, and XI).Binding of thrombin to thrombomodulin (receptor in endothelial cells) is more as compared to fibrin changes its state from procoagulant to anticoagulant (Al-Amer 2022).In the anticoagulant state thrombin binds with antithrombin.The lack of anticoagulant role of thrombin results in increased risk of venous thrombosis and haemophilia (Negrier et al. 2019).Antithrombin-III is the endogenous inhibitor of serine proteases involved in the coagulation cascade (Grover and Mackman 2022).Inactivation of thrombin by antithrombin-III results in the reduce production of fibrin (Cerskus et al. 1984).Thrombin being the main serine protease involved in clotting may serve as a promising target for the development of drugs for thrombosis (Govindarajan et al. 2016).The complex of antithrombin-III with thrombin interferes with the clotting process making antithrombin-III as the most significant physiological inhibitor involved in blood coagulation (Luxembourg et al. 2011).ATryn, antithrombin protein has been approved by the FDA as an anticoagulant for the prevention of clots before, during, or after surgery or birthing in patients (Salas and Miyares 2013).The risk of venous thromboembolism is 20 times higher in people with AT-III deficiency (Van Boven et al. 1999).Thus thrombin inhibitors from plants may be the novel alternative therapeutics (Kwon et al. 2004) for the development of plant derived anticoagulants.
In the present study we have isolated an antithrombin-III like protein from P. granatum.Based on sequence homology, docking and molecular dynamics studies ALPP showed the highest binding affinity with thrombin.Flexibility and stability of the complex was evaluated up to 50 ns.Preliminary validation of efficiency of thrombin inhibitor was done by performing in vitro thrombin inhibition studies.Results have indicated the presence of proteinaceous thrombin inhibitor(s) in P. granatum.This protein may be suggested and developed as a potential alternative phytotherapeutics against thrombosis.This is the first report of antithrombin-III like proteins from P. granatum.

Results and discussion
2.1.Antithrombin-III like protein from P. granatum (ALPP) and its interaction with thrombin Haemostatic balance between procoagulant (clot formation) and anticoagulant pathway (clot dissolution) is essential to maintain the fluidity of blood (Colman 2006).The imbalance between the procoagulant and anticoagulant pathways lead to the formation of thrombus in veins and arteries, thus resulting in 'thrombosis' (Palta et al. 2014).Thrombosis is a type of vascular disorder that causes morbidity and mortality (Raskob et al. 2014).Available anticoagulant drugs are associated with side effects thereby realizing the need for safer natural drugs (Ageno et al. 2012).Thus supplementing antithrombin-III with plant based antithrombin-III like molecule may serve as anticoagulant therapy.
If two proteins share 30% sequence identity their topologies may be the same (Radwan and Mahrous 2020).Thus sequences showing more than 32% homology with antithrombin-III in pBLAST search were shortlisted.The sequence of antithrombin-III (conserved regions from 402aa to 412aa) was aligned with the sequence of ALPP using EMBOSS Water local pairwise sequence to find the sequence homology in conserved regions and serpin domain (SP).Sequentially conserved residues of antithrombin-III (402F-412I) and ALPP (361F-372R) show sequence disparities only at one amino acid phenylalanine (F) where it forms loop and beta sheet respectively.Highly conserved regions (j) were represented by yellow colour (Figure S1).This shows the similarity of the sequentially conserved regions between the ALPP and antithrombin-III.
Modelled structure of ALPP passed the quality check parameters with favoured region (>90%) and showed <40% instability index.An instability index less than 40 using ProtParam (Walker 2005) tool classifies protein as stable (Shymaa 2014).In this case, ALPP instability index is computed to be 32.30and can be classified as stable protein.Ramachandran plot was generated using PROCHECK web server and it is expected that the most favoured region have over 90% residues, to be considered as a good model (Hooft et al. 1997).The ALPP has 92.4% residues in the most favoured regions (Figure S2) thus qualifying it as good quality model.
Antithrombin-III protein belongs to the serpins superfamily (Bjork and Olson 1997).Serpins are characterised by a common central region consists of 350 amino acids with 25-50% sequence homology to other classes of serpins (Sanrattana et al. 2019).The serpins contains a reactive centre loop (RCL), three beta sheets and 7-9 alpha helices (Sanrattana et al. 2019).The PyMol 2.4.0 visualized structure of antithrombin-III (Palta et al. 2014) and ALPP have also shown the presence of a reactive centre loop, beta sheets and nine alpha helices.This shows the structural similarity of ALPP with serpins.
ALPP and antithrombin-III structures were superimposed and analysed by the FATCAT server that provides access to protein structure alignment algorithm to visualize the differences between the protein conformation (Ye and Godzik 2004).Results indicated that the two superimposed structures are significantly similar with P-value of 6.02 e À12 and have 364 equivalent positions with RMSD of 3.44 Å and two twists with identity 26.13% and similarity of 43.47%.According to SuperPose server, in ALPP residues from 166L-175K forms beta-sheets connected by loops.However residues 188K-194L, 197G-204F are involved in the formation of antiparallel beta sheets.Another small size beta sheet is formed by residues from 213V-215A connected by a loop.In antithrombin-III, residues 113E-131Y, 156E-166Y are involved in formation of an alpha helix connected by loop.Residues 139K-149D forms beta sheets connected by loops.
Based on the docking scores and free binding energy, the protein sequence from P. granatum was shortlisted as antithrombin-like protein (ALPP).ALPP showed docking score À4828.59kcal/mol and free binding energy of À41.28 kcal/mol which is even better than control, thrombin-antithrombin-III docked complex which was À39.41 kcal/mol (Table S1).In another study for the identification of potential inhibitors of coagulation factor XIIa, the binding energy of docked eighteen compounds were in the range from À65.19 to À15.72 kcal/mol (Xu et al. 2020).
Inhibitors bind at the active site and exosites of thrombin in a competitive mode of inhibition.The beta sheet-like conformation of protease allows binding of the inhibitor at protease active site in a similar way as of substrate (Jmel et al. 2021).BBIs, Kazal and Kunitz domains type inhibitors are exosite assisted competitive inhibitors that binds to the secondary site (exosites) of proteases different from its active site because its active site gets blocked in a non-catalytic manner.The binding of inhibitor at exosite is known to enhance inhibitor specificity.Competitive protease inhibitors sometimes follow a modified type of inhibition mechanism where the N-terminal region of the inhibitor interacts with protease active site similarly as of substrate but gets shifted to side of the active site to ensure sufficient binding energy for the formation of enzyme-inhibitor complex (Mullan 2006).In our results, the ALPP interacts on either exosite I or exosite II of thrombin and act as a type of exosite assisted competitive inhibitor (Table S1).
2.3.Binding of ALPP to thrombin prevents the binding of substrate TAME to thrombin TAME (N a -p-Tosyl-L-arginine methyl ester hydrochloride) has been used as an esterolytic substrate for thrombin and binds at the active site residue Ser 195 (Sherry and Troll 1954).In molecular docking between thrombin and TAME using AutoDock Vina (Figure S6a), best conformation of TAME and thrombin shows binding affinity of À5.8 kcal/mol and binds at Ser 195 .The ALPP was docked with thrombin using AutoDock Vina (Figure S6b).No binding of TAME at residue Ser 195 of thrombin was found.This indicates that probably after binding with ALPP, thrombin loses its affinity to bind with TAME and this may also serve as the fact to validate our findings that protein isolated from P. granatum has the potential to act as inhibitor for thrombin.

Molecular dynamic simulation
Binding of ALPP at thrombin exosite prevents the binding of TAME at active site of thrombin, according to literature, inhibitors binding at the active site and exosites of thrombin generally follow competitive mode of inhibition (Jmel et al. 2021).Thus, ALPP may be characterized as following exosite assisted competitive mode of inhibition.It can be assumed that binding of TAME at active site provide more free energy as compared to its binding at other sites.Therefore, the substrate binding site was used for the molecular docking.The binding free energy of Thrombin-TAME was found to be À5.8 kcal/mol and the binding free energy of ALPP-Thrombin complex docked with TAME was found to be À7.2 kcal/mol.
TAME was docked at the active site of thrombin.The important interacting residues of thrombin obtained during molecular docking with TAME and their 3D interactions were observed (Figure S7).The flexibility and the stability of the Thrombin-TAME and ALPP-Thrombin complex docked with TAME were evaluated and confirmed using MD simulation for 50 ns.The average potential energy of the Thrombin-TAME was found to be À6.6 kcal/mol and for ALPP-Thrombin complex docked with TAME found to be À2.7 kcal/mol which remained energetically stable throughout the simulation (Figure S8a).
After 25 ns Thrombin-TAME complex, that is, positive control (black colour) becomes stable whereas ALPP-Thrombin complex docked with TAME (red colour) shows some level of fluctuations.The average Ca-backbone RMSD for Thrombin-TAME was found to be around 0.2 nm (Figure S8b) which is optimum and for ALPP-Thrombin-TAME complex it was found to be around 0.4 nm.The RMSF values of both the Thrombin-TAME and ALPP-Thrombin complex docked with TAME were as expected in the range of 0.3 Å (Figure S8c).Amino acids (171S, 259P, 260G, and 261F) of the ALPP-Thrombin complex docked with TAME fluctuated much more in the active site region at around 0.4 Å. MD simulation results showed distance between carboxyl oxygen of Ser 195 of thrombin and L-arginine of TAME is 1.2 Å.On contrary, the ligand undergoes orientation changes within the binding site of the thrombin.The TAME interacts with Ser 195 in the best-docked pose.Moreover, TAME was found to be interacting with 89Y, 132A, 133G, 134Y, 135K, 185K, 187D, and 260G of ALPP-Thrombin complex.
After performing 50 ns MD simulation of ALPP-Thrombin complex docked with TAME resulted in greater RMSD of ALPP-Thrombin and causes some structural changes by occupying the binding site of TAME on thrombin protein.These observations explain how ALPP-Thrombin complex disrupts the binding site of TAME and validate the inhibitory role of ALPP as an inhibitor.

In vitro assessment of thrombin inhibition activity in P. granatum
Presence of thrombin inhibition activity in P. granatum was also validated by performing in vitro thrombin inhibition assays with protein extracted from P. granatum leaves.Our preliminary in vitro experiments showed about 78% thrombin inhibition at 5 mg protein (Figure S9a).The IC 50 calculated for this crude protein fraction and positive control were 0.188 mg and 0.208 mg, respectively (Figure S9b).This indicates presence of proteinaceous thrombin inhibitor(s) in P. granatum.Earlier 77.5% protease inhibitor has been reported in flowers of P. granatum at 500 mg protein concentration (Rashid and Shafi 2018).SDS-PAGE profiling of crude protein (0-80%) showed protein bands of different intensity ranging from less than 15-250 kDa (Figure S10a).Reverse zymography of crude protein indicated the presence of thrombin inhibitors or thrombin resistant proteins (Figure S10b).This present work supports our in silico observations for the detection of antithrombin-III like protein in P. granatum.

Plant material
Fresh leaves of P. granatum were collected from campus garden of GGSIP University, Dwarka, New Delhi, India.Geographical coordinates of the sample collection site is 28 35 0 39.2382 00 N (Latitude), 77 1 0 12.5796 00 E (Longitude).A voucher specimen of P. granatum has been deposited at the herbarium of the University School of Biotechnology, Guru Gobind Singh Indraprastha University, New Delhi (voucher number: IPU-USBT/SJ/KKA/01/18).

Chemicals used:
Thrombin was procured from HiMedia, India, TAME (Na-p-Tosyl-L-arginine methyl ester hydrochloride) was purchased from Sigma-Aldrich.Other chemicals and reagents used in this study are of analytical grade.

Sequence and structure retrieval of antithrombin-III and thrombin
The full sequence of antithrombin-III, a natural thrombin inhibitor was retrieved in FASTA format from UniProt (UniProt ID: P01008).Antithrombin-III structure was downloaded in '.pdb' format from PDB database (PDB ID: 3EVJ) (Berman et al. 2000).Thrombin full-length sequence in FASTA format was retrieved from UniProt (UniProt ID P00734).Thrombin structure was downloaded in '.pdb' format from PDB database (PDB ID: 3GIC).InterProScan was used to analyse the sequence of antithrombin-III and thrombin (Blum et al. 2021).

Identification of interacting active site residues of thrombin with antithrombin-III
The thrombin (PDB ID: 3GIC) and antithrombin-III (PDB ID: 3EVJ) were docked using HawkDock server.The docked complex was analysed in PDBsum (Laskowski et al. 2018) server for all possible amino acid interactions and compared with PDB database available docked structure of Antithrombin-Thrombin complex (PDB ID:1TB6).
ERRAT for the measurement of connections among atoms (Colovos and Yeates 1993).PROCHECK web server was used to check the stereochemical quality of the proteins (Laskowski et al. 1993) and generates a Ramachandran plot showing the residues that are allowed in the most favourable region (Ramachandran and Sasisekharan 1968).WHATIF server provides an environment for homology modelling of protein tertiary and quaternary structures, validating protein structure, and correcting protein structure (Vriend 1990).VERIFY 3D analyses the compatibility of models with its amino acid sequences (1D) (Luthy et al. 1992).PROVE determines the volumes of atoms in macromolecule (Pontius et al. 1996).EMBOSS Water local pairwise sequence alignment helps in identifying the structurally or evolutionary similar regions between the two protein sequences (Shymaa 2014).Structural analysis was done using FATCAT and SuperPose server.Antithrombin-III and ALPP crystal structure were superimposed via SuperPose server to find the similarity in the arrangement of alpha helices, beta-hairpin, insertion or extended loops.SuperPose is a protein superposition server that helps to calculate superposition of proteins using a modified quaternion approach by generating sequence-structure alignment, RMSD statistics, difference distance plots and PDB coordinates (Maiti et al. 2004).
3.6.Molecular docking studies 3.6.1.Protein-protein interaction HawkDock server was used for docking modelled ALP structures with thrombin (PDB ID: 3GIC) (Hou et al. 2011;Feng et al. 2017;Weng et al. 2019).After initial preparations receptor and ligand files were prepared and uploaded on the HawkDock server.Thrombin was selected as ligand.Modelled shortlisted structures were selected as receptor molecules.After docking, MM/GBSA (Fu et al. 2018) analysis was done for the top 10 complexes using HawkDock server (Wang et al. 2019;Aljindan et al. 2021).The best docked modelled ALP structure was selected based on the rank of the docking score and binding energy expressed as kcal/mol.The best docked complex was analysed for possible amino acid interactions using PyMol (Schrodinger and DeLano 2020) and PDBsum.

Protein-ligand interaction studies
Thrombin-substrate (TAME) interaction: Molecular docking for Thrombin-TAME (substrate) complex was performed using AutoDock Vina tool.TAME was used as ligand and thrombin as the receptor.The crystal structure for TAME was downloaded from PubChem ID: 3083734 and was prepared for molecular docking.Thrombin (PDB ID: 3GIC) was used as a receptor and prepared before final docking studies.Docking results were analysed and visualized using PyMol 2.4.0.
Binding of TAME to Thrombin-ALP: Thrombin-ALP complex was docked with TAME to confirm whether the binding of ALP to thrombin blocks the binding of TAME to thrombin.Molecular Docking was performed using AutoDock Vina tool (Morris et al. 1998).TAME was used as ligand and Thrombin-ALP as the receptor.Structure for TAME was downloaded in '.sdf' format from PubChem ID: 3083734 and was prepared for docking in '.pdbqt' format.Thrombin-ALP were used as receptor and prepared before performing docking in '.pdbqt' format.Docking results were analysed and visualized using PyMol 2.4.0.

Molecular dynamic simulation
The flexibility and the stability of Thrombin-TAME complex and ALPP-Thrombin complex were evaluated as a function of time by performing 50 ns MD simulation using GROMACS 4.5.6.

In vitro detection of thrombin inhibition activity in crude protein from P. granatum
Fresh leaves (5 gm) of P. granatum were washed thoroughly with distilled water and homogenised by using extraction buffer.The homogenate was centrifuged at 9000 rpm at 4 C for 10 min and supernatant was collected as crude extracts.The collected supernatant was subjected to ammonium sulphate precipitation (0-80% saturation) (Wingfield 1998) by adding solid ammonium sulphate with continuous stirring at 4 C.After incubating it at 4 C for 30 min, the solution was centrifuged at 9000 rpm, 4 C for 10 min.The pellets collected and resuspended in minimum volume of buffer (10 mM Tris buffer, pH 7.2) and dialyzed overnight in dialysis buffer (10 mM Tris buffer, pH 7.2) at 4 C.After dialysis, sample was centrifuged at 9000 rpm at 4 C.The supernatant was collected and termed as crude protein.The protein concentration was estimated using Bradford assay (Bradford 1976) with BSA as a standard.

Thrombin inhibition studies
Thrombin inhibition activity of crude protein was determined by incubating thrombin (2 units) solution with different concentration (0.2, 0.5, 1, 2, 5, 10, and 20 lg) of crude protein separately at 25 C for 15 min (Bijak et al. 2014).After incubation, mixture was subjected to normal thrombin assay as mentioned below.Decrease in thrombin activity was used to calculate thrombin inhibition efficiency of crude protein.Thrombin activity without crude protein was used as control.

Thrombin assay
Thrombin (2 units) was added in 3 ml of 0.5 mM TAME (N a -p-Tosyl-L-arginine methyl ester hydrochloride)) in 50 mM Tris-Cl buffer, pH 7.2 in quartz cuvette and mixed properly.Change in absorbance was measured at 247 nm after every 60 s for 30 min.This was used as the control for thrombin inhibition assay (Sherry & Troll 1954).
3.8.3.Analysis of crude protein on SDS-PAGE (12.5%)Crude protein fraction was subjected to 12.5% SDS-PAGE.The gel was stained with coomassie brilliant blue (CBB) R-250 overnight and then destained to visualize the protein bands.
3.8.4.Reverse zymography of crude protein extracted from P. granatum leaves Crude protein was electrophoretically separated on SDS-PAGE (12.5%) copolymerized with gelatin (0.1%).After electrophoresis gel was washed in triton X-100 (2.5%) for 45 min with gentle shaking to remove SDS from the gel.Then gel was washed 3-4 times with distilled water and incubated for 24 h in developing solution containing thrombin (100 U) prepared in 50 mM Tris-Cl pH 7.2 and 10 mM CaCl 2 .Gel was stained with coomassie brilliant blue (CBB-0.25%)R-250 for 3-4 h and destained.Stained protein bands resistant to thrombin hydrolysis were observed on gel (Sharma & Bhattacharyya 2017).
3.8.5.Determination of IC 50 IC 50 value is calculated by plotting x-y values that fits the data in a straight line (linear regression) using this formulae: Y ¼ M Ã X þ C, where y ¼ 50.The IC 50 is defined as the inhibitor concentration at which half inhibition is obtained.

Conclusion
An antithrombin-III like protein (ALPP) has been identified from P. granatum.In silico experiments based on sequence homology, docking and molecular dynamics studies showed significant binding affinity of ALPP with thrombin indicating its potential as an inhibitor for thrombin.Detection of thrombin inhibition activity in the crude protein further confirms the presence of antithrombin-III like activity in P. granatum.More efforts are needed to purify and characterize thrombin inhibition activity for better understanding of its mode of action and utility as phytotherapeutics.