Screening of phytochemicals from Golden apple of discord, Cydonia oblonga against the pTen and HBx using in silico-based tools.

Background Phytochemicals derived from Cydonia oblonga have been investigated for their anti-oxidant and anti-cancer activities in many cancer cells lines. The reported bioactive compounds are evaluated in-silico to develop a novel antagonist against pTEN and HBx to target hepatocellular carcinoma. Lower expression of pTEN or higher expression of HBx represents the progression of hepatocellular carcinoma (HCC). This research is intended to identify the best candidate which interacts with our target proteins (pTEN and HBx) from the quince seeds by using computational methodologies. used in this investigation. The best docked result is compiled and structure were developed using the Discovery Studio Visualizer 3.5 and Pymol.


Introduction
Liver cancer or also known as Hepatocellular carcinoma (HCC), a tumor originates from hepatocytes is the disease which causes the second highest mortality worldwide according to the GLOBOCON report 2018. The major etiology agents accountable for HCC are consumption of a atoxin B1 in foods, DDT (Dichloro diphenyl trichloroethane) and organic chlorides in water and alcohol, chronic liver injury, alcoholic and biliary cirrhosis, hemochromatosis and due to virus infections like HBV and HCV (hepatitis B and C virus) [1]. Many studies have been reported that HCV infection was found in 40%-80% of HCC cases [2], that is strenuously correlated to liver insulin resistance which results in the causes of brosis and steatosis [3]. Several potential tumor suppressing genes (TSG) and oncogenes are used for the early-stage identi cation of HCC as a biomarker that is active in the progression of HCC.
Along with Arf, p53 and Ink4a, the tumor suppressor gene pTEN which is discovered in 1997 is one of the most in uential cancermutated genes. It is located speci cally from 87,863,438 base pare (b.p) to b.p 87,971,930 of chromosome 10 at position 23.3 and 10q23 locus. The crucial role of p10 is dephosphorylation of downstream phosphatidylinositol 3,4,5trisphosphate in normal hepatocytes, thus down-regulating signaling pathways like PI3K/Akt/mTOR [4,5]. pTEN, however, has negative regulative impact on downstream pathway in case of its deletion or mutation resulting a decline in the normal regulative mechanism of cell proliferation or metastasis, thereby causing the prevalence or malignant transformation of tumors. Presently, the abnormal behavior of pTEN is related to the incident and spread of various malignant tumors like endometrial, glioma, liver and prostate cancer [6]. Akt pathway suppression mediated by leads to increase in longevity, insulin signals down-regulation and resulting in tumor suppression [7]. The gene pTEN conjointly seems to be essential for in-vivo study for embryonic development in mouse [8].
Early growth response protein-1, Ras, ROS and NF-κB are some of the pTEN regulators [9]. In a data collected on transgenic mice, the core gene of HCV is believed to have a crucial role in HCC progression. by down-regulating the expression of retinoblastoma protein (Rb), and altering its tumor suppressive role [10]. The down-regulation of pTEN expression regulated by the core protein of HCV was reported in a research carried out by Zhang et al. (2014), which leads to the activation of downstream Akt gene at transcriptional level. In the in-vitro study, the HepG2 cells that are expressing the core protein of HCV shows signi cant enhanced proliferation and viability and cells had concentrated in G2/M cell cycle phase [11]. HBx mediated suppression of pTEN expression has also been reported to be involved in the HCC progression [12]. Additional studies have shown over-expression of Matrix Mettalo Protein-9 (MMP-9) as a result of Activator Protein-1 and NF-κB up-regulation and that are controlled through the ERK1/2 and PI3K pathway, leading to increased cell invasion in HBx transfected HCC [13]. pTEN role in HBx induced IGF-2 signaling via PKC or MAPK signaling was discovered to have a low level of IGF-2 at increased level of pTEN expression due to increased transcription factor Sp-1 phosphorylation, thereby limiting tumor progression [14].
The active use of the computational approach to drug development has been documented in several studies in recent years [15].
Computational study of the bioactive compounds of Graviola has been carried out for its anticancer activities, the study helped in narrowing down the number of compounds of interest and thus in reducing the cost of drug discovery [16] and many such studies have been done to screen the natural compounds for the particular target(s) or diseases [17] using in-silico approach. Cydonia oblonga commonly known as Quince has been reported for its broad spectrum of therapeutically values. They are used from centuries as a traditional medicine for constipation, diarrhea, gum problems, respiratory tract disorders, canker sores, asthma, bronchitis, sore throat, cough and even rhinitis. Many studies have been reported for their anti-oxidant, antiseptic, antiin ammatory, hepato-protective, and anticancer activities owing to the presence of therapeutically important bioactive compounds Therefore, this study attempts to index the reported naturally-derived bioactive compounds from the quince seeds for their anticancer activities especially those involve in traditional medicine systems are "drug-like" using the in-silico software's and AutoDock to specify the ADMET (Absorption, Distribution, Metabolism, Excretions and Toxicity) to determine pharmacokinetics pro le against cancer. Docking is carried out against pTEN and HBx proteins, as a target to evaluate the binding complex based on their binding score, residues in binding pockets and hydrogen bonds involved. Furthermore, the protein-ligand complex molecular dynamics simulation is conducted with the best binding score to ascertain the complex binding stability over a period of time.

The selected compounds study for Drug likeness parameters
The Lipinski rule of ve, known as Ro5 is used to estimate the permeability and solubility of the mentioned compounds which help in understanding their suitability for oral bioavailability [19]. The drug likeness was evaluated by utilizing an online server tool name Molinspiration (http://www.molinspiration.com). Ro5 predicts the poor permeation and absorption when the criteria are more likely: a). Greater than Ten Hydrogen Bond Acceptor (HBA). b). Molecular weight greater than 500kDa. c). More than 5 Hydrogen Bond Donor (HBD). d). Should be more than ve the octanol-water potential (XlogP) / Moriguchi log P (Mlog P>4.15). e). Antibody, vitamins, antifungal etc. are exceptional to this rule that belong to the classes of substrates for biological transporters.

ADMET pro ling of the bioactive compounds
The compounds satisfying the rule of ve are screened out and further investigated for their pharmacokinetics parameters. The criteria include the analysis of the properties of compounds like Toxicity, Absorption, Distribution, Metabolism and Excretion in the human system. The online server PreADMET is used for this parameter study (https://preadmet.bmdrc.kr/). Mutagenicity is predicted using the Ame's Test while carcinogenicity model is based on carcinogenic test on mice and rats for two years. The aspects we consider in our studies include parameters like skin permeability, Caco-2 permeability, mutagenicity, carcinogenicity, Human Intestinal Absorption (% HIA), MDCK cell permeability, cytochrome P450 2D6 metabolism and blood-brain barrier penetration.

Target protein structure preparation
The three-dimensional structure of the protein pTEN (1D5R) and HBx (3MS6) are retrieved from the RCSB protein database (Research Collaboratory for Structural Bioinformatics) (www.rcsb.org). BIOVIA Discovery Studio tool is used to prepare the target in such a way that it has no ambiguities i.e., small molecules and water molecules are deleted. The ligand/substrate attached to the target protein and heteroatoms are removed. The structure is energy minimized to remove bad steric clashes using CHARMM force eld. In this study, all the computation studies were conducted in vacuum phase.

Preparation of ligand structure
Seventeen bioactive compounds are selected in these studies, which are reported to be present in the C. oblonga seeds in various review literatures. The compounds structures are retrieved from the PubChem compound database and later converted to its 3D PDB structure via Discovery Studio 2.5 software where it was energy minimized. Furthermore, gasteiger partial charges were added and CHARMM force eld and MMFF94x partial charge were applied to optimize the structures. Using the RMS gradient energy, structure is further minimized with 0.001 kcal/mol, although other parameters are maintained by default.

Molecular docking of ligands with pTEN (TSG) and HBx oncogene
AutoDock tools (ADT) is one of the most cited in-silico tools in research eld used for performing the docking experiment. This software uses the AMBER force eld that include Genetic Algorithm and scoring function of Empirical free energy to explore the binding a nity between the target protein and ligand compound [20]. AutoDock is employed to execute docking simulations, generating ten conformations of ligand in complex with the target receptor, which were further ltered through their free binding energy score. The protein structure is parameterized by allocating atomic salvation parameters, Kollman charges, polar Hydrogen atoms, and fragmental volume while ligands were energy minimized using ADT. Four coordinate les are created that are ligand.pdbqt, receptor.pdbqt, grid.gpf and dock.dpf. The co-ordinates of x, y and z axis of each grid point was established around 60 x 60 x 60Å and for pTEN the co-ordinates were positioned at X= 38.248, Y=82.41 and Z=31.742, while for HBx it was set at X=0.724, Y= 9.825 and Z= 9.071 with a grid spacing between 2 connecting points was kept at 0.375Å of the grid. The molecular docking simulation for 150 population size were conducted using energy evaluations at 2500000 for a number of hundred GA (Genetic algorithm) for each ligand and RMSD (Root Mean Square Deviation) for each evaluation for docking was set at 2.0Å with 0.274 co-e cient of torsion degree of freedom were constitute for inhibitor molecules.
The molecular docking was accomplished using the Lamarckian genetic algorithm (LGA) and the remaining parameters were set to default with a 0.2 rate of mutation, 0.8 crossover rate with 27,000 numbers of maximum generations. The results were generated in .glg and .dlg le were further studied for the ligand and protein interaction. The results were clustered based on their binding energy, inhibition constant and intermolecular energy [21]. The vary protocol was performed on each ligand used in this investigation. The best docked result is compiled and structure were developed using the Discovery Studio Visualizer 3.5 and Pymol.
2.6. Stability study of the protein-ligand complex by molecular dynamic simulation The ligands having highest free binding energy when docked against pTEN and HBx were further continued for its stability analysis using GROMACS 5.1 package software (GROningen MAchine for Chemical Simulations) [22]. SwissParam, the online server was utilized to create the ligand topology [23] and receptor protein topology was developed with the aid of the CHARMM27 all-atom force eld within the framework of GROMACS functions. A water model of 3.0 nm distance TIP-3 points (TIP3P) to sustain the simulation under periodic boundary conditions was introduced from the protein to the box faces [24]. A dodecahedral box model with SPC water molecules was used for solvation and to prevent clashes of proteins, a margin of 10 Å was applied with external boundaries [25]. In all simulation process, the solvated system is neutralized by adding sodium ions. To bring the maximum force below 1000 kJ/mol/nm about 2000 steps of steepest energy minimization was executed for removing the steric clashes for all the systems. The electrostatic interactions with long-range were evaluated by employing the method of Particle Mesh Ewald (PME) and the bond length was constraint by applying the LINCS (Linear constraint solver) algorithm [26], respectively. For the computation of the non-bonded interaction of Van-der Waals forces, applied cut-off of about 1.0 nm value [27]. Further, our system is equilibrated for 1000 picosecond under NVT at temperature of 300K (constant number of particles, volume, and temperature) followed by 1000 picoseconds NPT equilibration at a temperature of 300 K with a pressure of 1 bar. The temperature coupling was accomplished at a temperature of 300K with a time constant of 0.1 picosecond by implementing velocity rescaling thermostat [28]. The algorithm provided by the Parrinello-Rahman barostat was implemented with a 1 picosecond constant to calculate the pressure at 1 bar [29].
Finally, the protein-ligand system was introduced to simulation of 100 ns of molecular dynamics (MD), performed under identical conditions at 1 bar and temperature of 300K. At each step, a Leap-frog integrator was applied throughout the MD simulation with a size of 2 femtoseconds and the resulted MD coordinates for further analysis were saved at an interval of every 2 picosecond [30]. The gmx rms, gmx rms, gmx sasa, gmx hbond and gmx gyrate module of GROMACS is used during the 100 ns long simulation to determine RMSD, RMSF (Root mean square deviation and Root Mean Square Fluctuation, respectively), solvent accessible surface area, number of hydrogen bonds involved and change in compactness using radius of gyration of protein backbone with reference to the atom's located at the initial positions in the backbone.

Results
Drug likeness and bioactivity score The concept "drug-likeness" is emphasized in pharmaceutical research in recent years as it's believed that successful prediction of drug-like properties in early stage will pay off in the drug development at the later stages. The Lipinski rule of ve, "Ro5" is the most favorable technique to identify the compounds that are similar to the conventional drug. If the violation is 0 or 1, the compounds are assumed to bind effectively to the receptor and was rejected when the violation number exceeds 2. Seventeen compounds that are present in C. oblonga seeds are selected (Table 1) for this study and are ltered through Lipinski rule, including the reference compound Doxorubicin. Table 1 In silico-based determination of pharmacokinetics of the selected compounds and Doxorubicin (Standard compound) using online server Molinspiration. Only 5 of these compounds including 3CQA (3-O-caffeoylquinic acid), Catechin, 4CQA (4-O-caffeoylquinic acid), Chrysoeriol and 5CQA (5-O-caffeoylquinic acid) are screened which follows the Lipinski drug likeness criteria and hence they are continued for further evaluations. Compounds that debased the Lipinski's rules more than once could have a bioavailability complication.
Two of our ltered compounds Chrysoeriol and Catechin show better bioavability in comparison to the standard drug Doxorubicin.
The drug is expected to bind the biological targets; thus, bioactivity of the compounds contributes to the overall potential to qualify the compounds as a drug that are determined by the combination of kinase inhibitor, G-protein-coupled receptor, protease and enzyme inhibitor, ion channel modulators and nuclear receptor ligands. According to the general concept, greater the bioactivity scores the probability of being biologically active of the individual compound is higher ( Table 2). If the bioactivity score for the metal complexes is higher than 0.0, then the complex is believed to be biologically active and if its value lies between -5.0 and 0.0, then the complex is active moderately while the compounds trusted to be inactive when bioactivity score lies below -5.0 [31]. 3CQA, 4CQA and 5CQA acids shows highly bioactive towards the nuclear receptor ligand. All our compounds show moderate activity toward other biological receptors (>-0.50). The results demonstrate that 3CQA and 5CQA shows better bioactivity score than doxorubicin. Table 2 Bioactivity score of the bioactive compounds and its complex evaluated using online server Molinspiration.

Sr.
No. Unwanted ADMET properties of the new drug candidate are one of the reasons for many developmental failures hence to survive the phase I clinical trials these problems required to be identify in earlier stages of processes. The hurdle encounter in traditional ADME and toxicity testing is (multistep and time-consuming) is addressed by employing combinatorial chemistry and computation studies. Hence for the determination of pharmacokinetics or pharmacology properties of the compound, several chemical parameters were evaluated to determine their compliance with the standard range by using in silico tools. Those properties comprise of blood-brain barrier penetration, aqueous solubility, MDCK cell permeability, Human Intestinal Absorption (% HIA), Toxicity, Caco-2 and skin permeability and Cytochrome P450 2D6 binding. These properties were projected to be essential for the invention of novel bioactive compounds, thus on investigate the failure of lead competitors that will cause toxicity in its inactive form or one can possibly penetrate the intestinal layer. Thus, on the basis of the evaluation our compounds show mutagenicity and moderate absorption/permeation properties, though doxorubicin is non-mutagenic but is carcinogenic in both mouse and rat. CQA shows non-carcinogenicity in mouse and carcinogenicity in rat (except 4-O-caffeoylquinic acid is non carcinogenic in both rat and mice), whereas catechin is carcinogenic in both mouse and rats ( and Chrysoeriol shows the promising results and are further evaluated for docking and its stability. Table 3 The ADMET profiling of the bioactive compounds and the reference compound (doxorubicin) using online server tool PreADMET.  The residues that are reported to be present at the active site in pTEN are HCXXGXXR motifs, while the residue Cys-124 and Arg-130 plays a critical role in pTEN catalytic activity and residue His-123 and Gly-127 are essential for the conformation P-lop of pTEN [32,33]. Additionally, Asp-92 residue present in "WPD" loop of pTEN behaves as basic acid to promote the phenolic oxygen protonation of the tyrosyl leaving group [34]. On the other hand, the binding pocket of HBx in the novel structure 3MS6 contain residues Leu-5, His-8, Thr-12, Thr-12, Thr-36, His-41, Glu-68, Ser-69, Asp-70, Asn-71, Ile-74, His-79, Asp-80 and His-87. The detailed evaluation of the binding pockets of caffeoylquinic acid derivatives with both pTEN and HBx is given in Table 5. As we compare the binding pocket of our docked results ( Fig. 1 and Fig. 2) the residues involved in Doxorubicin are ALA126, ASP92, ARG130, CYS124, GLY127, GLY129, HIS93, ILE168, LYS125, LYS128, LYS330, THR167 and similar residues are seen in other ligands binding pockets ( Fig. 1) against docking with protein pTEN. Hence this shows that are ligands are binding in the vary pocket where the novel pTEN (1D5R) structure reported to have active sites. Our standard complex with pTEN shows 5 hydrogen bond interactions while caffeoylquinic acid derivatives shows hydrogen bonding more than Doxorubicin-pTEN complex, which also attributes to the stability of the complex. Table 5 Active site pocket residues and Hydrogen bonds interaction between protein pTEN (1D5R) and ligand revealed  ASP92, HIS93, CYS124,   LYS125, ALA126, GLY127,   LYS128, GLY129, ARG130,   THR131, THR167, ILE168,   GLN171, LYS330   HIS8, LEU9, THR12,   VAL21, ILE74,   ILE76, GLN77,   HIS79, VAL84, ALA85, VAL86,

LYS88
The pocket residues involve in HBx-doxorubicin complex (Table 5)  Hence it can be concluded that the binding of the ligands is in the similar pocket region as our standard drugs doxorubicin is binding with the HBx.
Molecular dynamics simulation to analyze the complex stability Initially, our docking results revealed that the screened compounds including 3CQA, 4CQA, 5CQA and Chrysoeriol could be potential inhibitors. 3CQA and 5CQA were, however, seen to be promising. Therefore, we employed MD simulation for better comprehending the molecular mechanism which may be involved in the analysis. An unbiased simulation was performed for each of the two protein-ligand complexes. A valuable insight was provided by the 100 ns long MD simulation related to the pTEN and HBx structural dynamics with different ligand compounds. For example, the energy of interaction between proteins and ligands illustrates the profound binding between them and well-built interactions between them are characterized by the higher number of hydrogen bonds.
MD trajectories were analyzed for all the related atoms in the backbone and ligands for RMSD determination and assessment of average uctuations i.e., the RMSF of each residue of the protein in a time-dependent manner for all molecules in the backbone.
One of the key criteria for testing the equilibrium of MD directions is the RMSD. To begin with, we determined the RMSD from the initial docked position of the individual compound. The smaller the variance, the individual compounds have a better docked conformation. In other words, during computation, the ligand remains in its initial conformation. The average RMSD value of pTEN with 3CQA and 5CQA for 100 ns trajectory calculated to be 0.245 nm and 0.215 nm, respectively. At rst, there is a small uctuation in the trajectory but later the degree of uctuation reaches equilibrium (Fig. 3 (a)). The pTEN complexes with 3CQA shows with average RMSF value about 0.100 nm. The residue at position between 4093 to 4115shows the highest uctuation in all the pTEN complexes. In comparison with standard, pTEN-5CQA shows more stability than pTEN-3CQA and pTEN-Dox as they have less variation throughout the 100 ns simulation and the system reaches equilibrium after 10 ns as compared to pTEN-Dox approximately and pTEN-3CQA. However, the system pTEN-5CQA shows higher exibility followed by pTEN-3CQA complex and lastly pTEN-Dox complex system (Fig. 3 (c)).
The standard-HBx complex system have average RMSD value about 0.376 nm, which reaches equilibrium after 30 ns. However, HBx-Dox complex shows higher deviation and lowest exibility compare to HBx-3CQA and HBx-5CQA complex system Fig. 3 complex system with the highest deviation. The RMSF graph of HBx with ligands shows an increase in uctuation compare to Doxorubicin, which might be due to the ligand adjustment in the binding pocket. Whereas, HBx-3CQA shows higher uctuation in residues than to HBx-5CQA complex at the active region compare to the standard RMSF.

Protein compactness and Hydrogen bond analysis
The compactness and the folding behavior of the protein during the simulation is analyzed by utilizing the structural parameters of the 3D protein for computing the Radius of gyration (Rg), which provides insight to the protein packaging during the simulation time period [35]. which is possibly due to the exposure of some surface residues. However, it immediately reached equilibrium after 1ns without any shift in structure during the whole run, depicting the protein stability before and after the ligand binding (Fig. 4 (c) & (d)).
To validate the complex stability the intramolecular bonds of hydrogens also play a key role [37]. The intramolecular hydrogen bonds within 0.35 nm dynamics were determined in protein pTEN and HBx. The number of hydrogen bonds in pTEN with Dox, 3CQA and 5CQA had maximum 236, 238 and 238, respectively and that of HBx complexes were computed to be maximum around 61, 66 and 67, respectively. The pTEN-ligand and Hbx-ligand complex shows a slight increase in intramolecular bonds compare to the Doxorubicin complex of both proteins, which is suspected due to the protein's higher compactness (Fig. 5 (a) & (b)). Similarly, the intermolecular hydrogen bonds show the similar number of bonds supporting the molecular docking analysis. The number of bonds involve in protein and compounds binding pocket to be around 4 to 6 for all complexes. Hence, supporting higher stability of the ligand complexes in comparison to the standard drug Doxorubicin throughout the trajectory (Fig. 5 (c) & (d)). Hence, the docking and dynamics analysis of the hydrogen bonds depict that binding of 3-CQA and 5CQA with pTEN and HBx is more signi cantly stable than that with the protein-Dox as there are least variation in system throughout the 100 ns simulation.

Discussion
The plant kingdom contains high levels of phenols, which are known for their biological activities (like antioxidant, anti-diabetic, antibacterial, antiviral, anticancer and many more). Although, their health bene ts reported from decades, studies concerning including biochemistry, analytical chemistry, biomedicine, etc. of these compounds are limited. Within this framework, many researches are attracted towards the un-investigated plant species to nd novel compounds to combat diseases including cancer, diabetes mellitus and Alzheimer. The plants that are especially used in folklore medicines are extensively studied and identi ed for their bioactive compounds having selective apoptosis inducing activity and growth arrest of cancerous cells with minimum side effects. For example, cinnamic acid derivatives, ferulic acids, caffeic acid is stated to modulate signaling pathways for survival and proliferation, such as MAPK and PI3K-Akt [38]. Hepatotoxicity caused by drugs and alcohol consumption are the critical reason for liver damage. The most common drugs including amoxicillin-clavulanate, azathioprine and in iximab are the implicated agent resulting in drug-induced liver injury. Amoxicillin-clavulanate mediated liver injury was observed to occur in one out of 2300 users in latest prospective research, approximately [39].
Unlike other cancers, many oncogenes and TSG are engaged in hepatocarcinogenesis [40] in which pTEN, p53, TGF-α, IGF-2 and Myc are some genes that are altered in HCC [41]. A study conducted in China, lower pTEN expression are associated with grade of tumor and stage of disease and metastasis which correlates with enlarged tumor size and higher mortality rate [42]. Therefore, to de ne the stage of cancer and to ascertain the prognosis of HCC in patients, the level of pTEN expression can be exploited. HBx plays a central role in the development of tumorigenesis; HBV-associated liver disease is involved in the transition from chronic hepatitis and cirrhosis to HCC through stimulating intercellular signaling pathways. HBx mainly mediates by interacting with constituents in the liver tumor microenvironment including liver cancer cells, hepatic stellate cells, in ammatory cytokines, immune cells, exosomes and HIF-1 α [43]. In HCC advancement, overt participation of HBx oncogene has also been implicated.
HBx directly disrupts the p53-mediated transcription of pTEN [44]. About 36% of HCC patients show higher expression of HBx protein; this can be utilized as a useful marker in HCC diagnosis. Activation of pTEN or suppression of HBx may be considered for a targeted therapy as an active approach against human HCC by the bioactive compounds.
Thus, for this purpose in-silico tools were utilized in selecting and forecasting the bioactive compounds of C. oblonga for their HCC potential. Quince seed have been studies for its therapeutically bioactive compounds which include 3CQA, 4CQA, 5CQA, diCQA, apigenin and chrysoeriol derivatives, stellarin-2 and leucenin-2 [45]. The AutoDock Tools 4.0 has successfully screened 3CQA and 5CQA bioactive compounds of C. oblonga by targeting pTEN and HBx proteins, before performing any expensive wet laboratory studies, recommending them as a potent anti-cancer agent for HCC management.
Several studies have been conducted to discover the biological attributes of C. oblonga and it derived bioactive compound caffeoylquinic acid derivatives, both in-vitro and in-vivo. An in-vivo experiment demonstrated about the protective role of aqueous extract of quince fruit in diethylnitrosamine (DEN) induced HCC rats, was reported by Adiban and his colleagues (2016) which shows remarkable reduction in serum biomarkers cancer and liver damage [46]. Caffeoylquinic acid (CQA) and its derivatives consumption stated to enhance antioxidant status, might increase colon motility and decrease colorectal cancer risk as they exhibit many physiological functions such as radical scavenging activities, anti-cancerous and antiproliferative activities [47]. The quince fruit methanolic extract (seed, pulp and peel) were analyzed for their antioxidant activity, shows signi cant free radical scavenging potential and antioxidant activities that organic acid fraction which is correlated with the caffeoylquinic acid content [48]. Similar study was conducted, to show the ability of isolated phenolics compounds from the methanolic extract of quince fruits, seeds and leaves to quench the DPPH (2,2'-diphenyl-1picrylhydrazyl) free radicals and to inhibit the oxidative hemolysis results by 2,2'-azobis(2-amidinopropane) dihydrochloride in human erythrocytes [49] and comparatively, the seed extract shows the strongest antioxidant activity. Quince Aqueous Fermented Extract (QAFE) shows an effective anion superoxide radical (ID50 = 73.7 µg/ml) and DPPH free radicals (ID50= 68.80 µg/ml) scavenging ability than Quince Wax Extract (QWE) which is comparatively more effective in preventing thiobarbituric reactive species (ID50= 48.9 µg/ml) [50].
The docking result outcome of the 3CQA and 5CQA shows some promising results for targeting pTEN and HBx protein as a potent approach for targeted therapy. According to the Lipinski rule of ve, 3CQA and 5CQA makes it a drug-like molecules except the negative value of Milog P i.e., poor solubility and permeability. However, in comparison with the standard drug doxorubicin it is quite similar (mentioned in Table 1). The ADMET pro ling of 3CQA and 5CQA compounds Caco permeation, MDCK, BBB, HIA and Skin permeability was in harmony with the reference drug Doxorubicin. 3CQA and 5CQA shows mutagenicity and carcinogenic in rat model but non carcinogenic in mouse. However, the toxicity of reference drugs shows carcinogenicity in both mouse and rats but non-mutagenic in Ame's test. As we know all mutagen agents are not necessarily carcinogens, we consider that factor and continued the further docking studies. Previous experimental studies have already highlighted the activity of chlorogenic acid and its derivatives as an effective therapeutic compound for MMP-2 and MMP-9 suppression and prevents HCC prognosis by inactivation of ERK1/2 pathway [51].
The crystal structure of both pTEN and HBx was retrieved from RCSB protein data bank to perform the docking using AutoDock  [42]. Subsequently, this computational study suggests that the C. oblonga phytochemicals especially 3CQA and 5CQA might show effective binding a nity with the target proteins pTEN and HBx respectively, as explained via AutoDock scoring function [52].
The 100 ns long MD simulation analysis disclosed that the complex stared 5CQA showing complex reaching equilibriumy after about 10 ns approximately and the ligand-pTEN complex systems show higher exibility than standard-protein complex system but ligand-HBx system shows exibility lower than the HBx-Dox as the uctuations are quite higher comparatively to standard.
Similarly, the Rg plot of p10 shows least deviation compare to HBx complexes, depicting the higher stability of p10 complexes than HBx complexes.

Conclusion
In developing countries about 70-95% of population depends on the traditional medicines and in market nearly 20-25% of modern drugs are derived from higher plants either directly or indirectly [53]. This attracts the scientist to exploit the traditional medicine system for cancer care. For example, "PHY906" is a Chinese herbal formulation which was approved in 2013 for phase II clinical trial in human [54]. From this study, it is clear that caffeoylquinic acid complex with pTEN shows signi cant binding a nity than with HBx protein target. In addition, incorporating Lipinski's drug likeness rule, bioactivity score, pharmacokinetics pro ling and also molecular docking suggests that CQA plays an important therapeutic role compared to conventional drug doxorubicin.
Therefore, 3CQA and 5CQA which also present in higher concentration in quince seed were taken for further molecular dynamic simulation. The simulation analysis based upon the RMSD and average hydrogen bonding a rms the stability of our complex 3CQA-pTEN, 5CQA-pTEN, 3CQA-HBx, and 5CQA-HBx that the standard drug doxorubicin. From these computational studies a conclusion is drawn that ligand 3CQA and 5CQA interacts with pTEN and HBx a rming to its anti-cancerous activity which are reported in many in-vitro studies mentioned above. Hopefully, a supportive hypothesis of developing HCC inhibitors will be put forward by the introduced compound. Thus, it might be a useful candidate for liver cancer prevention and care.
Declarations Figure 1 The best docked complex structure of ligands with pTEN tumor suppressor gene

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.