Catechin and curcumin interact with S protein of SARS-CoV2 and ACE2 of human cell membrane: insights from computational studies

The recent outbreak of the coronavirus (SARS-CoV2) is an unprecedented threat to human health and society across the globe. In this context, development of suitable interventions is the need of the hour. The viral spike protein (S Protein) and the cognate host cell receptor ACE2 can be considered as effective and appropriate targets for interventions. It is evident from the present computational study, that catechin and curcumin, not only exhibit strong binding affinity to viral S Protein and host receptor ACE2 but also to their complex (receptor-binding domain (RBD) of the spike protein of SARS-CoV2 and ACE2; RBD/ACE2-complex). The binding affinity values of catechin and curcumin for the S protein, ACE2 and RBD/ACE2-complex are − 10.5 and − 7.9 kcal/mol; − 8.9 and − 7.8 kcal/mol; and − 9.1 and − 7.6 kcal/mol, respectively. Curcumin directly binds to the receptor binding domain (RBD) of viral S Protein. Molecular simulation study over a period of 100 ns further substantiates that such interaction within RBD site of S Protein occurs during 40–100 ns out of 100 ns simulation trajectory. Contrary to this, catechin binds with amino acid residues present near the RBD site of S Protein and causes fluctuation in the amino acid residues of the RBD and its near proximity. Both catechin and curcumin bind the interface of ‘RBD/ACE2-complex’ and intervene in causing fluctuation of the alpha helices and beta-strands of the protein complex. Protein–protein interaction studies in presence of curcumin or catechin also corroborate the above findings suggesting the efficacy of these two polyphenols in hindering the formation of S Protein-ACE2 complex. In conclusion, this computational study for the first time predicts the possibility of above two polyphenols for therapeutic strategy against SARS-CoV2.

www.nature.com/scientificreports/ promotes the membrane fusion process inside the host cell 9,10 . Recently, it has been reported that receptor binding domain of S Protein of SARS-CoV2 is more or less similar to that of SARS-CoV, despite amino acid variation at some key residues 2 . This suggests that the virus can also target ACE2, a monomeric membrane bound protein of human cells 11,12 . Therefore, it is presumed that ACE2, the cognate receptor of corona virus present on the cell membrane of host cells can also be a specific target to prevent the viral entry 13 . Several recent studies have suggested that natural polyphenolic compounds like catechins (GTCs; Green Tea Catechins) and curcumin (diferuloylmethane; from turmeric) have antiviral activities against a broad spectrum of viruses such as Human Immunodeficiency Virus (HIV), Herpes Simplex Virus, Influenza Virus, Hepatitis B and C Viruses (HBV and HCV respectively) 14 , Adenovirus 15 and Chikungunya virus (CHIKV) 16 . Diverse mechanisms have been suggested to explain the antiviral activities of both the polyphenolic compounds. For example, GTCs have been documented to be a potential suppressor of viral entry and its replication [17][18][19][20][21] , while curcumin has been demonstrated as a potent inhibitor of monophosphate dehydrogenase, a rate limiting enzyme in the de novo synthesis of guanine nucleotide 22 . Further, it has also been observed that GTCs and curcumin inhibit the expression of ACE2, as evident from animal studies 23,24 .
Although catechin and curcumin have been reported to bind with various proteins of viral and human origin, there is not much information on interaction of polyphenols with S protein of the coronavirus and its cognate receptor, ACE2 of host cells till date. With this back drop, the present study has been designed to examine interaction of catechin and curcumin with S protein of the virus and its cognate receptor ACE2 of host cells employing computational methods. Computational approaches (Molecular docking and simulation) are the first and foremost choice of scientists to prophesize apparent binding modes and affinities of ligands for macromolecules before experimental studies which are indeed expensive and time consuming 25 . In addition, improvement in speed, reliability and accuracy of computational docking methods in last few years have made it a suitable choice to design structure-based drugs 26,27 . The present study incorporates results of molecular docking of catechin and curcumin with the S Protein of corona virus as well as with ACE2 of host cells, a cognate receptor for viral S Protein. In addition, binding affinities and molecular simulation studies of catechin and curcumin with the 'RBD/ ACE2-complex' indicate that both the polyphenols cause considerable alteration in the structure of the complex.   33 . This minimisation of structures is done using OPLS3e force field. Each input structure generates multiple output structures due to different stereochemistry, protonation states, tautomer's and ring conformations. In the ligand output file specifications are made for production of one low energy ring conformation per ligand. Grid-based Ligand Docking with Energetics (GLIDE) module in Schrodinger software was used for the formation of S Proteincurcumin and S Protein-catechin complex. Also catechin and curcumin were individually complexed with the 'RBD/ACE2-complex' . Desmond software was used for carrying out molecular dynamics simulations, Root Mean Square Deviations (RMSD) and atomic fluctuation through Root Mean Square Fluctuation (RMSF) studies. For conducting explicit solvent simulations with periodic marginal conditions, different tools such as cubic, orthorhombic, truncated octahedron, rhombic dodecahedron and other arbitrary simulation boxes are used. Prior to 100 ns production run, 8-staged stabilization run was conducted which includes primarily task, then simulations in Brownian Dynamics with NVT at T = 10 K, small time steps, restraints on solute heavy atoms for 100 ps and followed by repetition of the above stage but with restraints on solute heavy atoms for 12 ps. The stage 4 was carried out in a similar manner to the previous one at NPT instead of NVT followed by focus on solvate pocket. The stage 6 is the same as that of stage 4. The next stage involved simulation at NPT for 24 ps with no restraints. Finally, simulations were done.
Here in this study, MD simulations were conducted notably for the top two identified hits to analyse the stability of the ligand receptor complex for 100 ns. Stability of docked complexes 2019-nCoV spike glycoprotein-curcumin and 2019-nCoV spike glycoprotein-catechin are simulated till 100 ns simulation time by performing Molecular Dynamics (MD) simulations using system builder of Desmond 34 implemented in Maestro 35 with OPLS3e force field. Simulations were also conducted for both the ligands with 'RBD/ACE2-complex' . The system for 'S Protein-curcumin' and 'S Protein-catechin' were immersed in a water filled cubic box of 10 Å spacing containing 63,985 (approximately) water molecules with system builder of the Desmond in the Maestro program. Similarly, for 'RBD/ACE2-complex' with catechin and curcumin approximately 26,850 water molecules were taken using extended simple point charge (SPC). Neutralisation of the docked complex was done by the addition of 4 Na + ions (1.137 mM concentration) into the system for S Protein and curcumin. 6 Na + ions (1.706 mM concentration) were added for neutralisation of S Protein and catechin. Similarly, 23 Na + ions (15.575 mM concentration) and 21 Na + ions (14.206 mM concentration) were added for neutralisation of 'RBD/ACE2-complex' for catechin and curcumin, respectively. Molecular Mechanics Generalized Born Surface Area (MM-GBSA) method has been adopted for the calculation of binding free energies of catechin and curcumin with S Protein and 'RBD/ACE2-complex' , respectively. The more negative value indicates stronger binding as the MM-GBSA is an index of free energy of binding. Prime module 36  www.nature.com/scientificreports/ Protein complexes revealed that the system is stable. Analysis of different conformations acquired over the simulation period of 100 ns is done. For the computation of average change in the displacement of selected atoms in a particular frame with respect to reference frame, Root mean square deviation (RMSD) is estimated for the protein and ligand for 100 ns simulation trajectory. In order to understand the unbinding trends of both proteins, chain A (Angiotensin-converting enzyme 2) and chain B (RBD of Spike glycoprotein) and to analyse the consistency of such trends in presence of catechin or curcumin, we carried out non-bonding interaction qualitative analysis by slicing every 5 ns frames from the Molecular dynamics trajectory of 100 ns.

Results and discussion
Structural analysis. Prediction of secondary structure S Protein of SARS-CoV2 has been done using SOPMA (Self Optimised Prediction Method with Alignment). The S Protein contains 1288 aa residues comprising 350 α helices (27.17%), 312 β-turns (9.08%) and 509 random coils (39.52%). Through ExPASy ProtParam, the total number of negatively charged (Asp + Glu) and positively charged residues (Arg + Lys) were determined to be 112 and 100, respectively. The aliphatic index was found to be 81.58. The GRAVY (Grand Average of Hydropathicity) scored to − 0.163. The instability index was computed to be 31.58. These features classify S Protein of SARS-CoV2 as a stable structure . It was also revealed through computational studies that the half-life of S Protein is maximum in case of mammals (mammalian reticulocytes-30 h) than yeast (> 20 h) and bacteria (E. Coli − > 10 h).
Structure alignment. SARS-CoV2 and SARS-CoV were evaluated by TM-align (https ://zhang lab.ccmb. med.umich .edu/TM-align /) for comparative structural studies. These two viruses were considered for Structure-Structure superimposition due to maximum sequence similarity. It was observed through structural alignment studies that SARS-CoV2 and SARS-CoV only differ in RBD fragment and remaining part of the structure is identical (Supplementary Fig. S2). It was apparent that SARS-CoV is an ancestor of the newly upsurging virus SARS-CoV2. Nevertheless, some changes were noticed in the RBD fragment of SARS-CoV2 compared to SARS-CoV. The results corroborate an earlier study 2 .
Protein-protein docking. Based on the total RMSD value the best 10 docking models with different free energies were obtained from the ClusPro web-server. Out of which, we analysed 5 ClusPro docking models which were selected based on probability of S Protein, S Protein with curcumin and S Protein with catechin to interact with the predicted binding sites of ACE2 with lowest binding energy during such interactions. Average binding energy of all 5 binding positions for S Protein-ACE2 interaction in the absence of curcumin or catechin is − 901.2 kJ/mol ( Supplementary Fig. S16). Nevertheless, average binding energy for S Protein-ACE2 in presence of either of the above two polyphenols is − 759.54 kJ/mol (Table 1, Figs. 2, 3, 4).
It was observed that during protein-protein interaction binding energy of S Protein-ACE2 decreases in presence of the phytocompounds, (i.e. curcumin or catechin). A significant decline in 141.66 kJ/mol of binding energy was observed during the interaction of S Protein-ACE2 in presence of curcumin or catechin compared to their direct binding. Therefore, it can be presumed that both the compounds are capable of hindering the attachment of RBD site of S Protein to the ACE2 receptor protein. This would indeed pave a way for the utilisation of curcumin or catechin in repurposing/design of effective therapy to prevent the viral entry.  Table 2). The binding affinity of curcumin with ACE2 was noted to be − 7.8 kcal/mol where as that of catechin was found to be − 8.9 kcal/mol (Table 3). Similarly, catechin and curcumin have − 9.1 kcal/mol and − 7.6 kcal/mol binding affinities, respectively, towards the 'RBD/ACE2-complex' ( ACE2 and also 'RBD/ACE2-complex' are higher than that of curcumin. The results from Molecular Simulation data throw a light on the interaction of curcumin with S Protein. The Root Mean Square Deviation (RMSD) of S Protein-curcumin complex was marked to increase for the initial 20 ns and then remained stable up to 100 ns during the simulation trajectory ( Supplementary Fig. S5). Local changes along the protein chain were characterised through Root Mean Square Fluctuation (RMSF). The plot indicates curcumin possesses the ability to cause fluctuation of all amino acids of S protein ( Supplementary  Fig. S6). Table S1 represents S Protein and ligand interactions which clearly depict the resident time of specific amino acid residues participating during the process, including the amino acids in the RBD site which display considerable interaction with curcumin. Keto group of curcumin exhibits a high affinity to Leu335 of RBD site of S Protein forming hydrophobic bonds. Interaction with Leu335 of RBD site of S Protein occurs for 40% of the simulation time ( Supplementary Fig. S7). Molecular simulation studies are in good agreement with docking studies. Results suggest that both polyphenols bind to S Protein with high binding energy, however, their binding sites on S Protein differ considerably. Curcumin binds directly to the RBD of S Protein whereas catechin binds to proximity of RBD of S Protein. In addition, catechin causes greater fluctuation in amino acid residues near the RBD site.
The RBD fragment of SARS-CoV2 spans from 319-591 S-residues 37 . From our studies it is deduced that curcumin directly binds to amino acids in this region Leu546, Gly548, Phe541, Asp571, Ala570, Thr572, Thr547, Thr573 whereas, catechin binds to the S Protein in the near proximity of RBD fragment to Gln314, Glu309, Lys310, Gly311, Lys304, Tyr313, Thr302, Ile312, Leu303 and Ile312 residues ( The average change in displacement of atoms in all frames was recorded through Root Mean Square Deviation (RMSD) at 10 ns interval. Although the maximum MM-GBSA binding energy for S Protein-curcumin complex was observed − 58 kcal/mol at 0 and 20 ns, the minimum value was recorded − 47 kcal/mol at 10 and 70 ns. On the other hand, the highest and the lowest MM-GBSA binding energy for S Protein-catechin complex were noticed to be − 59 kcal/mol (at 30 ns) and − 20 kcal/mol (at 60 and 80 ns), respectively (Supplementary Fig. S14). Total www.nature.com/scientificreports/ MM-GBSA free energy calculation for both polyphenols with S Protein indicates a favourable binding energy for the curcumin (− 53.63 kcal/mol) as compared to catechin (− 34.22 kcal/mol). Average RMSD of both complexes (S Protein-catechin complex and S Protein-curcumin complex) was recorded less than 3 Å after stabilization of S Protein RMSD. On the contrary, RMSD of S Protein-catechin complex was initially unstable for the first 50 ns and stabilised thereafter for the rest of the simulation period ( Supplementary Fig. S11). Maximum structural fluctuation was observed in between 300 and 500 amino acid residues and after 1000 amino acids residues of S Protein (Supplementary Fig. S10). The above data supports that S Protein and catechin interaction occurs with amino acid residues of S Protein near the RBD site (319 aa-591 aa) 37 . Amino acid residues Arg634 and Val635 near the RBD site of S Protein have stronger affinity towards hydroxyl group of catechin with 54% and 35%, respectively, out of 100 ns simulation trajectory (Supplementary Fig. S12). The binding affinity of curcumin with ACE2 was noted to be − 7.8 kcal/mol where as that of catechin was found to be − 8.9 kcal/mol. The binding of curcumin or catechin with ACE2 includes conventional hydrogen Bond, carbon-hydrogen bond and Pi-Sigma interactions. The amino acid residues of the protein that take part in the above interactions vary for both ligands (Supplementary Figs. S4, S9 and Table 3).
The binding affinity of 'RBD/ACE2-complex' with curcumin and catechin scored to be − 7.6 kcal/mol and − 9.1 kcal/mol, respectively. Results of the present investigation suggest that amino acid residues of both the components of 'RBD/ACE2-complex' that interact with catechin and curcumin are different. While curcumin binds with the ACE-2 receptor in the complex through van der Waals interactions (Phe390, Asn33, Glu37, His34, Asp38, Lys353, Arg396, Pro389), catechin interacts with ACE2 receptor in the complex through van der Waals interactions (Arg559, Gln338, Ser536, Thr92, Leu29, Val93, Lys26, Asn33, Ala386 and His34), conventional hydrogen bonds (Arg393, Glu37, Gln96), carbon-hydrogen bond (Ala387), pi-alkyl bond (Pro389) and pi donor hydrogen bond (Gln96). Curcumin is also engaged with the S Protein segment of the complex through van der Waals interactions (Arg408, Gly416, Ile418, Tyr505, Gln493, Gly496, Tyr495, Arg403), conventional hydrogen bonds (Gln409, Ser494) (Table 4), carbon-hydrogen bond (Glu406); pi-alkyl bond (Lys417), unfavourable donor-donor (Gln409) and pi donor hydrogen bond (Tyr453). The interaction of catechin with the S Protein region of the complex involves van der Waals interaction (Gly416, Gln409, Asp405, Leu455, Tyr453), conventional hydrogen bonds (Glu406, Arg408, Tyr505) and pi-alkyl bond (Lys417) ( Table 4). The number of www.nature.com/scientificreports/ hydrogen bonds formed during interaction of curcumin and catechin with 'RBD/ACE2-complex' are 3 and 7, respectively. Such H-bonds are formed between the amino acid residues of the protein complex and OH groups of the polyphenols. Catechin shows high binding affinity towards the receptor due to presence of more number of functional OH groups. The average RMSD of the protein complex (S Protein and ACE2) with catechin and curcumin was calculated as 1.8 Å (Fig. 7), and 2.0 Å (Fig. 8), respectively. The extent of fluctuation caused in the protein complex due to binding of catechin and curcumin was 3.3 Å and 4.2 Å, respectively. Further,  www.nature.com/scientificreports/ these two polyphenols are also able to cause fluctuation in both alpha-helical and beta-strand regions of the complex. While curcumin-mediated fluctuation is more localized to S Protein part of the complex than ACE2, catechin-mediated fluctuation is observed in both the components of the complex (Fig. 9). The maximum and the minimum MM-GBSA binding energies for 'RBD/ACE2-complex' with curcumin were found to be − 69 kcal/ mol (at 10 ns) and − 53 kcal/mol (at 90 ns), respectively. Similarly, during interaction of 'RBD/ACE2-complex' with catechin, the maximum and the minimum MM-GBSA binding energies were found to be − 57 kcal/mol (at 20 ns) and − 28 kcal/mol (at 80 ns), respectively ( Supplementary Fig. S15). Total MM-GBSA free energy calculation indicates a favourable binding energy for the interaction of curcumin with 'S Protein and ACE2 complex' (− 60.84 kcal/mol) as compared to that of catechin (− 45.04 kcal/mol). Upon docking of both polyphenols on the interface of 'RBD/ACE2-complex' concomitant with 100 ns MD simulation, it was portrayed that stable and favourable interactions are formed by the polyphenols with both RBD and ACE2 proteins. Further, a higher binding affinity of both polyphenols was observed towards RBD than ACE2.
Results of the present study suggest that amino acid residues of both the components of 'RBD/ACE2-complex' that interact with polyphenols used in the present study considerably differ depending on the nature of polyphenols as well as the components of the complex. In addition, interaction time of amino acid residues of both the components of the complex varies depending upon the types of polyphenol. For example, the residues like Table 3. The binding energy, types of interaction and amino acids involved in the interaction of human ACE2 receptor with curcumin and catechin.
Results obtained from multiple evidences depicted that both the polyphenols bind preferably to sites of S Protein (RBD site) which are crucial in host cell binding 33 . Similarly, it was also seen that these molecules attach to those sites of ACE2 which were involved in serving a medium of viral entry 39 . Thus, results of the present computational studies suggest possible prevention of the viral infection by the use of catechin and curcumin, two widely used natural polyphenols. This dual inhibitory machinery of blocking the binding of host cell receptors to virus and inhibiting cellular entry of viral protein could be an effective therapeutic target, as evident from an array of computational studies. However, this needs to be experimentally validated prior to translational intervention.
In addition, elimination and neutralization of viral infection by catechin and curcumin cannot be ignored because both the polyphenols are well acclaimed immuno-stimulant and inducer of autophagy, another important mechanism of viral clearance 14,40 .
Therefore, availability of catechin and curcumin in the host system may facilitate all different mechanisms simultaneously and, thereby promote elimination and/or neutralisation of viral infection.

Conclusion
The pandemic novel corona virus has created a stark landscape in the social, health and economic sphere. The lethality of the virus has taken many lives. There is urgency to curb the widespread outbreak of SARS-CoV2. In this context, findings of this computational study indicate that catechin and curcumin can be considered for prospective antiviral drugs against SARS-CoV2. Nevertheless, this requires further experimental validation to substantiate the findings.  www.nature.com/scientificreports/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.